Microbiology (2008), 154, 3686–3696 DOI 10.1099/mic.0.

2008/021584-0

Divergent roles of CprK paralogues from

Desulfitobacterium hafniense in activating gene
expression
Krisztina Gábor,3 Kagnew Hailesellasse Sene, Hauke Smidt,
Willem M. de Vos and John van der Oost
Correspondence Wageningen University and Research Centre, Laboratory of Microbiology, Dreijenplein 10,
Krisztina Gábor Wageningen 6703-HB, The Netherlands
krisztina.gabor@codexis.com

Gene duplication and horizontal gene transfer play an important role in the evolution of prokaryotic
genomes. We have investigated the role of three CprK paralogues from the cAMP receptor
protein–fumarate and nitrate reduction regulator (CRP–FNR) family of transcriptional regulators
that are encoded in the genome of Desulfitobacterium hafniense DCB-2 and possibly regulate
expression of genes involved in the energy-conserving terminal reduction of organohalides
(halorespiration). The results from in vivo and in vitro promoter probe assays show that two
regulators (CprK1 and CprK2) have an at least partially overlapping effector specificity, with
preference for ortho-chlorophenols, while meta-chlorophenols proved to be effectors for CprK4.
The presence of a potential transposase-encoding gene in the vicinity of the cprK genes indicates
that their redundancy is probably caused by mobile genetic elements. The CprK paralogues
activated transcription from promoters containing a 14 bp inverted repeat (dehalobox) that closely
resembles the FNR-box. We found a strong negative correlation between the rate of
transcriptional activation and the number of nucleotide changes from the optimal dehalobox
Received 23 June 2008 sequence (TTAAT-N4-ATTAA). Transcription was initiated by CprK4 from a promoter that is
Revised 23 August 2008 situated upstream of a gene encoding a methyl-accepting chemotaxis protein. This might be the
Accepted 26 August 2008 first indication of taxis of an anaerobic bacterium to halogenated aromatic compounds.

INTRODUCTION N-terminal cysteine motif for iron–sulphur cluster binding,

hierarchically controls one of the three DNR homologues
Members of the cAMP receptor protein–fumarate and
that lack the cysteine-based sensory module (Vollack et al.,
nitrate reduction regulator (CRP–FNR) family of tran-
1999). Another example of a regulatory cascade mediated
scriptional regulators show exceptional diversity in their
by CRP–FNR homologues is found in Rhodopseudomonas
effector specificity and the promoters they target. As such,
palustris (Egland & Harwood, 2000). In this facultatively
they control a wide range of physiological processes, anaerobic a-proteobacterium, the degradation of 4-hydro-
ranging from nitrate and fumarate respiration (FNR) to xybenzoate is regulated by the CRP–FNR-type HbaR,
glucose starvation by sensing cAMP levels (CRP) (Bauer which is hierarchically controlled by the oxygen-sensing
et al., 1999; Kolb et al., 1993). At present, over 350 AadR protein, also from the same family.
members of the CRP–FNR family have been identified,
with many examples of more than one CRP–FNR Due to the growing number of sequenced bacterial
homologue in the same organism (Korner et al., 2003). genomes, the known CRP–FNR family has expanded
The relationship between these related CRP–FNR tran- rapidly (Korner et al., 2003). Genome analysis of
scriptional regulators is often complex, showing hierarch- Escherichia coli has identified the third member of the
ical characteristics. As an example, the facultative anaerobic CRP–FNR family, YeiL (Anjum et al., 2000). Apart from
c-proteobacterium Pseudomonas stutzeri possesses four Bradyrhizobium japonicum and Magnetospirillum magneto-
FNR homologues, of which FnrA, which contains the tacticum, the chromosome of Desulfitobacterium hafniense
DCB-2 shows the highest diversity of CRP–FNR-type
3Present address: Codexis Laboratories Hungary Kft., 1045 Budapest, regulators encoded in a single bacterium (Mesa et al.,
Berlini utca 47–49, Hungary. 2006). D. hafniense belongs to the lineage of Gram-positive
Abbreviations: CHPA, 3-chloro-4-hydroxyphenylacetic acid; CRP, cAMP
low-G+C bacteria, and its most important characteristic is
receptor protein; DCP, dichlorophenol; EMSA, electrophoretic mobility the capability for anaerobic respiration with halogenated
shift assay; FNR, fumarate and nitrate reduction regulator; HTH, helix– compounds as terminal electron acceptors (halorespira-
turn–helix; RNAP, RNA polymerase; TCP, trichlorophenol. tion) (Christiansen & Ahring, 1996). This strictly anaerobic

3686 2008/021584 G 2008 SGM Printed in Great Britain

 Transcriptional activation by CprK paralogues

bacterium can couple the reductive dehalogenation of ively, are: ZP_01372871 and Dhaf_0678 (CprK1), ZP_01372893 and
often toxic meta- and ortho-substituted phenol derivatives Dhaf_0698 (CprK2), ZP_01372888 and Dhaf_0693 (CprK3),
to its growth, thereby providing a potential means of ZP_01372914 and Dhaf_0718 (CprK4), and ZP_01372902 and
Dhaf_0707 (CprK5). Conserved protein domains were identified
bioremediation of polluted anoxic environmental sites with the help of the Pfam database (http://www.sanger.ac.uk/
(Van Eekert & Schraa, 2001). The key enzymes in Software/Pfam), and HNN secondary structure prediction tool
halorespiration are the reductive dehalogenases (RDs), (http://npsa-pbil.ibcp.fr). The multiple sequence alignment was
corrinoid/iron–sulphur-containing proteins that are pre- produced by CLUSTAL_X 1.81 (Thompson et al., 1997) and edited in
dicted to be membrane anchored by a small protein, the GeneDoc 2.6 (Nicholas et al., 1997). The consensus dehalobox
gene for which generally clusters with the reductive sequence was obtained using the WebLogo 3 BETA program (http://
dehalogenase-encoding gene (Smidt & de Vos, 2004). threeplusone.com/weblogo/) (Schneider & Stephens, 1990). The raw
data output of the same program was used to obtain a weight measure
From D. hafniense, DCB-2, a halorespiration-inducible
(conservation value) for each position of the dehalobox consensus.
ortho-chlorophenol-reductive dehalogenase (CprA1) has The raw data output (given as bit/symbol) was converted to units of
been isolated and characterized (Christiansen et al., 1998). the graphical output data (bits) by multiplying the values by 1/ln2.
The transcription of cprA1 is activated by CprK1, a novel Correlation coefficients were computed in Excel (Microsoft Office
member of the CRP–FNR family, in the presence of 3- 2003).
chloro-4-hydroxyphenylacetic acid (CHPA) (Gábor et al.,
2006). It has been shown that binding of CHPA to CprK1 Identification of putative regulator-binding DNA sequences.
results in an active DNA-binding conformation which The training datasets for the motif discovery MEME algorithm (Bailey
& Elkan, 1994) consisted of 40 DNA sequences with an average size of
enables the regulator to activate transcription from 150 bp from the region directly upstream of the translational start
promoters that contain a specific DNA sequence termed codon of halorespiration genes. As a positive control, sequences that
the dehalobox (TTAAT-N4-ATTAA) (Mazon et al., 2007). contained known FNR-like regulatory boxes from the D. dehalogenans
The partially assembled sequence data for the genome of D. chlorophenol-reductive dehalogenase (cpr) gene cluster were included
hafniense DCB-2 reveals the presence of at least 20 proteins in the analysis (Smidt et al., 2000). Additionally, manual detection of
that fulfil selection criteria for the CRP–FNR family, inverted repeats was performed with the Palindrome tool (http://
namely (i) a length of 230–250 aa, (ii) a C-terminal helix– mobyle.pasteur.fr/cgi-bin/MobylePortal/portal.py?form=palindrome).
turn–helix (HTH) DNA-binding motif, and (iii) an N-
Overproduction and purification of CprK-like proteins. CprK1
terminal nucleotide (effector)-binding domain (Korner et was prepared as described previously (Gábor et al., 2006). The genes
al., 2003). Five of the CRP–FNR family members encoding CprK2 and CprK4 were amplified from D. hafniense
(including CprK1) show high similarity (36–89 % identity genomic DNA using oligonucleotide primers, as listed in Table 1.
at the amino acid level) with CprK from Desulfitobacterium After digestion with the appropriate endonucleases, PCR products
dehalogenans (Pop et al., 2004) and cluster with genes that were cloned into linearized pET24d expression vectors (Novagen).
encode potential halorespiration proteins (Villemur et al., The resulting T7-based expression vectors containing one of the cprK
2002). We have focused on these putative transcriptional homologues (Table 2) were introduced into E. coli JM109(DE3) cells
by heat-shock transformation. Overproduction of the recombinant
regulators in our research, while other CRP–FNR family
proteins was done essentially as described for CprK1 (Gábor et al.,
members that do not cluster with halorespiration genes 2006), with the exception that IPTG-induced CprK2 and CprK4
have been left for further investigations. The unusually production was carried out at 37 instead of 20 uC. Purification of
high occurrence of CprK-like regulators is likely to be CprK2 and CprK4 was performed similarly to that of CprK1, using
correlated with the relatively large number of halogenated sequential HiPrep heparin and Superdex 200 (Amersham Biosciences)
compounds that this organism can accept as terminal chromatography columns (Gábor et al., 2006).
electron acceptors (Madsen & Licht, 1992), enabling a
In vivo promoter probe assays. DNA fragments carrying potential
specific response by each regulator to a specific group of
promoter elements and CprK-binding sites (dehaloboxes) were PCR-
halogenated compounds. amplified from D. hafniense genomic DNA and digested by
The aim of the present paper was to investigate the role of endonucleases, followed by ligation with linearized pAK80 promoter
multiple CprK paralogues in D. hafniense by using in vivo probe vectors. The oligonucleotide primers and the resulting
promoter probe vectors are listed in Tables 1 and 2, respectively. E.
promoter probe assays as well as in vitro DNA-binding coli JM109(DE3) cells were co-transformed with the pAK80
assays. Our results indicate that at least two of the five derivatives and the corresponding expression vectors carrying a
CprK paralogues have a distinct effector range, which most cprK homologue, as indicated: pWUR226 with pWUR216, pWUR218
likely reflects the gain of a new or specific function by or pWUR219; and pWUR227 with pWUR220, pWUR221, pWUR222
divergence of redundant genes. or pWUR223. Selection pressure during aerobic growth in Luria–
Bertani medium was maintained by the addition of 30 mg kanamycin
ml21 (pET24d derivatives) and 200 mg erythromycin ml21 (pAK80
derivatives). Triplicate experiments were carried out for each
METHODS condition, and b-galactosidase activity was measured as described
Bioinformatics and statistical tools. The genome sequence of D. previously (Gábor et al., 2006). Throughout the experiments, 1 Miller
hafniense DCB-2 was screened for halorespiration genes using basic unit was defined as follows: 10006A420/(t6v6OD600), where A420 is
local alignment search tool (BLAST) (McGinnis & Madden, 2004). absorbance at 420 nm, t is reaction time, v is sample volume and
Protein accession numbers of the identified CprK homologues (based OD600 is the optical density of the culture at the time the sample was
on D. hafniense genome sequence version 19-Jun-2006) and Dhaf taken. In control experiments, recombinant cells carrying the pET24d
locus tags from the IMG database (http://img.jgi.doe.gov), respect- vector instead of pWUR226 or pWUR227 were used.

http://mic.sgmjournals.org 3687
K. Gábor and others

Table 1. Oligonucleotide primers used in the study

Primer Sequence* OrientationD

BG1385 CGCGCCCATGGAAAGGGTAATAAGCAATCAC S
BG1391 GCGCGCCATGGGAGAAATTCTTAAAAATTATATTTTTCC S
BG1748 GGTTGAGAAATTCAGGTAAAG S
BG1749 GGATCACATACGCAAGTATTAATG A
BG1826 CGCGCAAGCTTGCTATGATGTTTATCTTCCTCC S
BG1827 CGCGCGGATCCGGACGGCATCCTTCCTTTG A
BG1828 GCGCGAAGCTTGGCCGGTCTTGTTGCCC S
BG1830 CGCGCGGATCCGCTTCACCAGGAAAGGCGG A
BG1831 GCGCGAAGCTTGCTCTTCTCGAAGAGGGGATAGC S
BG1832 CGCGCGGATCCCAGGATAGTTCCCATTTTTTCACC A
BG1835 GCGCGAAGCTTCCGAGGTTGAGAGCTTTAATTG S
BG1836 CGCGCGGATCCGTACTCATCCCTTTCACCTCC A
BG1838 GCGCGAAGCTTGGCCCCTAATTTATGGAG S
BG1839 CGCGCGGATCCGCTAAAAACACAATCCCGGAC A
BG1840 GCGCGAAGCTTCCTGAAGAGAGCCCTTGATC S
BG1841 CGCGCGGATCCGGAATGAGCGGATCTTGAATG A
BG1842 GCGCGAAGCTTGCAACCAGCGGCTTCGCC S
BG1843 CGCGCGGATCCGATATGGATTGCACTAAGTTCCC A
BG1940 GCGCAAGCTTAATAAGCTATCCCCTCTTCG A
BG1941 GCGCGGATCCTCAGAATCTCAATTCCTCTTCC A
BG2109 GCTCATTTTCCAAATTGGCG S
BG2110 GGTACCAGAATAGTATAAAG A

*Introduced endonuclease restriction sites are underlined.

DS, sense primer; A, antisense primer.

Electrophoretic mobility shift assay (EMSA). The 85 bp dsDNA 50 mM Tris/HCl, pH 7.5, 5 mM MgCl2, 2.5 mM EDTA, 250 mM
fragment containing the DB8 dehalobox was PCR-amplified from D. NaCl), 2.5 mM DTT, 1 mg poly(dG-dC)-poly(dG-dC), 0.5 mM
hafniense genomic DNA using oligonucleotide primers BG2109 and purified CprK1 or 1 mM purified CprK4, and 1.5–2 nM 32P-labelled
BG2110. The 52 bp PCR product containing DB7 was obtained using DNA. Additionally, potential effector molecules were added from a
BG1748 and BG1749 primers. DNA fragments were purified 2.7 mM aqueous stock solution, to a final effector to protein molar
according to the modified ‘crush and soak’ method and 59-labelled ratio of 800 : 1. The reaction mixtures were first incubated at 24 uC for
using [c-32P]ATP as described previously (Gábor et al., 2006). EMSA 30 min to allow complex formation, then loaded on a 6 %
reaction mixtures (20 ml) contained 16POP buffer (20 % glycerol, polyacrylamide gel buffered with 89 mM Tris and 89 mM boric acid

Table 2. Plasmids used in this study

Plasmid Description Primers Source or reference

pET24d Expression vector (5.3 kb), pMB1 ori, KanR, IPTG-inducible T7 promoter – Novagen
pAK80 Promoter probe shuttle vector (11.0 kb), p15A/L. lactis ori, EryR, – Israelsen et al. (1995)
promoterless lacLM genes
pWUR166 DB7:lacLM promoter fusion containing pAK80 derivative BG1704/1743 Gábor et al. (2006)
pWUR168 DB6:lacLM promoter fusion containing pAK80 derivative BG1702/1742 Gábor et al. (2006)
pWUR171 DB5:lacLM promoter fusion containing pAK80 derivative BG1699/1782 Gábor et al. (2006)
pWUR176 cprK1 gene cloned in pET24d BG1379/1814 Gábor et al. (2006)
pWUR216 DB8:lacLM promoter fusion containing pAK80 derivative BG1826/1827 This study
pWUR218 DB9:lacLM promoter fusion containing pAK80 derivative BG1828/1830 This study
pWUR219 DB10:lacLM promoter fusion containing pAK80 derivative BG1831/1832 This study
pWUR220 DB1:lacLM promoter fusion containing pAK80 derivative BG1835/1836 This study
pWUR221 DB2:lacLM promoter fusion containing pAK80 derivative BG1838/1839 This study
pWUR222 DB3:lacLM promoter fusion containing pAK80 derivative BG1840/1841 This study
pWUR223 DB4:lacLM promoter fusion containing pAK80 derivative BG1842/1843 This study
pWUR226 cprK4 gene cloned in pET24d BG1391/1940 This study
pWUR227 cprK2 gene cloned in pET24d BG1385/1941 This study

3688 Microbiology 154

 Transcriptional activation by CprK paralogues

(~pH 8.1) and electrophoresed at 10 mA constant current at 4 uC, Transcriptional activation by CprK paralogues in
followed by drying and autoradiography. Band intensities were E. coli
quantified with GeneTool 3.08 (Syngene) software using the manual
baseline correction method. In order to study the nature of the putative CprK-binding
motifs (dehaloboxes) that were found upstream of the cpr
genes (Fig. 1a, b), a recombinant in vivo promoter probe
RESULTS assay was developed in E. coli. Each cprK-encoding gene
was PCR-amplified from D. hafniense genomic DNA,
CprK-like regulators and target sequences in cloned into the pET24d expression vector and used to
D. hafniense transform E. coli JM109(DE3) host cells. After optimization
of the culture conditions for protein synthesis, CprK1
The genome of the halorespiring D. hafniense DCB-2 (26.5 kDa), CprK2 (26.5 kDa) and CprK4 (27 kDa) were
contains five ORFs (cprK1–cprK5) that share 36–89 % successfully overproduced in E. coli, partly in the soluble
amino acid sequence identity with the transcriptional fraction (Fig. 1c). Despite repeated efforts, the soluble
regulator CprK from D. dehalogenans (Smidt et al., 2000). overproduction of CprK3 and CprK5 was not achieved.
A BLAST search also identified several clustering genes that, CprK1 appeared as a double band during SDS-PAGE
similarly to the previously annotated cprTKZEBACD analysis (Fig. 1c); MS measurements have demonstrated
cluster of D. dehalogenans (Smidt et al., 2000), were that this is (at least in part) due to removal of the N-
predicted to encode the following: chlorophenol reductive terminal methionine (Mazon et al., 2007). A promoterless
dehalogenases (cprA), small hydrophobic membrane b-galactosidase-encoding gene was fused with D. hafniense
anchors for the reductive dehalogenase (cprB), GroEL-type promoter fragments, each containing a dehalobox centred
chaperones (cprD and cprE), proline cis/trans isomerases in the middle of a 0.2 kb DNA fragment. The resulting
(cprT), and putative transcriptional regulators from the promoter probe constructs (pWUR216–223) are listed in
NosR/NirI family (cprC) (Fig. 1a). One of the gene clusters Table 2.
contains a putative transposase-encoding gene from the IS4
family (trn) that is next to a truncated reductive We tested the in vivo transcriptional activation within gene
dehalogenase gene (rdhD) and a Tat-signal peptide- clusters by co-transforming E. coli JM109(DE3) host cells
encoding ORF (tat). All the five D. hafniense CprK with the CprK2-encoding plasmid (pWUR227) and
paralogues were predicted to contain an N-terminal dehaloboxes from its own gene cluster (DB1–DB4), or
effector-binding domain (Pfam no. PF00027; cNMP- the CprK4-encoding plasmid (pWUR226) and dehalo-
binding domain) and a conserved C-terminal HTH boxes DB8–DB10. Triplicate cultures of each promoter
DNA-binding motif (Pfam no. PF00325). Indeed, the fusion-containing strain were grown under two conditions:
recently solved crystal structure of CprK1 from D. in the presence of 20 mM CHPA as effector compound, or
hafniense confirmed the presence of these domains (Joyce in the absence of the halogenated compound. The
et al., 2006). functionality of CprK2 and CprK4 on the respective
promoters was analysed by detecting the activity of the
In the recognition a-helixes (aF) of the HTH domain,
b-galactosidase reporter enzyme (Fig. 3), and data were
CprK1 and CprK2 possess a conserved motif V--SR (Fig.
compared with results that had previously been obtained
2); in CprK4 the corresponding sequence is V--SK, whereas
for CprK1 on DB5–DB7 (Gábor et al., 2006).
in CprK3 and CprK5 it is V--CK. The position of these
residues corresponds to the motif E--SR in E. coli FNR, CprK2 was most active on the promoter that contains
which has been demonstrated to be responsible for the dehalobox DB1 and is situated upstream of the putative
specificity towards its target DNA sequence TTGAT-N4- reductive dehalogenase-encoding genes cprBA2 (Fig. 3a).
ATCAA (FNR-box) (Green et al., 2001). Similar activity was measured on the promoter of cprK3,
containing DB3. The putative promoters that contain DB2
Based on the high similarity of the CprK paralogues to
and DB4 dehaloboxes were weaker, showing approximately
characterized CprK transcriptional regulators as well as on
40 % residual activity.
their chromosomal position, a search was performed for
regulator-binding motifs in the upstream sequences of all Remarkably, transcriptional activation by CprK4 was only
the cpr genes. Ten putative regulatory sequences, termed observed with DB8, a dehalobox that is situated upstream
dehaloboxes (DB1–10), were identified using the MEME of a putative methyl-accepting protein-encoding gene
algorithm (Bailey & Elkan, 1994); in addition, one distantly (macA) (Fig. 3c). The fact that the promoter of macA
related motif was found in the promoter region of the encoding an important player in the chemotaxis machinery
putative 3,5-dichlorophenol (3,5-DCP) reductive dehalo- is recognized by a CprK homologue might suggest that D.
genase-encoding gene cprBA5 (DB0) (Fig. 1a, b). The hafniense elicits chemotactic responses to chlorinated
identified sequences included three dehaloboxes that have compounds. Promoter activity mediated by CprK4 from
been studied in detail previously (Gábor et al., 2006). Each DB9 or DB10 (upstream of cprK4 and cprBA4, respectively)
dehalobox contains a 5 bp imperfect inverted repeat with was negligible. This indicates that the transcription of
4 nt spacing, and the consensus sequence of dehaloboxes is cprK4 and cprBA4 might be activated by a different
TTAGT-N4-ACTAA (Fig. 1b). regulator, or alternatively, that their transcription can be

http://mic.sgmjournals.org 3689
K. Gábor and others

Fig. 1. Identification and production of CprK homologues. (a) Reductive dehalogenase gene clusters in D. hafniense DCB-2.
Arrows indicate putative CprK-binding sites (dehaloboxes). (b) Sequence alignment of dehaloboxes. Arrows are placed above
the 5 bp inverted repeats; the 39 nucleotides correspond to putative Pribnow boxes. Dehalobox and Pribnow box consensus
sequences generated by the WebLogo 3 BETA program are shown to the right of the alignment. (c) SDS-PAGE gel of the
soluble (1) and insoluble (2) cell fractions of recombinant E. coli JM109(DE3) cells overproducing CprK1 (26.5 kDa), CprK2
(26.5 kDa) or CprK4 (27 kDa). Full-size heterologously produced proteins are indicated by the black arrowhead; the grey
arrowhead corresponds to the truncated CprK1 protein.

initiated by CprK4 from macA by readthrough. One Dehaloboxes: perfect fit on a perfect repeat?
important difference was observed in the pattern of
transcriptional activation by CprK-like proteins: while In vivo promoter probe experiments showed that, similarly
CprK1 (Fig. 3b) and CprK2 (Fig. 3a) were strongly to CprK1, the two new CprK paralogues (CprK2 and
regulated by CHPA, showing a drastically reduced CprK4) are also involved in the recruitment of RNA
promoter activity when this compound was absent from polymerase (RNAP) and consequently in transcriptional
the media, CprK4 was equally active with or without activation in E. coli (Fig. 3). All the D. hafniense promoter
CHPA (Fig. 3c). We have studied this phenomenon further fragments that were used in the assay harbour a 14 bp
using in vitro EMSA experiments. inverted repeat, termed the dehalobox, as a potential

3690 Microbiology 154

 Transcriptional activation by CprK paralogues

Fig. 2. Multiple alignment of the protein sequence of selected members of the CRP–FNR family. The secondary structure of
CprK1 is included below the alignment (based on PDB 2H6B, monomer A); a-helixes are denoted by cylinders, b-sheets by
arrows. The cylinders corresponding to the characteristic HTH DNA-binding motif are in black. Residues that are involved in
CHPA binding in CprK1 are indicated by diamonds, while asterisks highlight amino acids that confer DNA-binding specificity to
FNR and to the CprK homologues. CRP, cAMP-binding protein from E. coli; FNR, fumarate and nitrate respiration regulator
from E. coli; CprK, transcriptional regulator of halorespiration from D. dehalogenans; CprK1–K5, known or putative
transcriptional regulators of halorespiration from D. hafniense.

binding target for the CprK paralogues. CHPA-induced difference in protein concentration does not explain the
promoter activities varied by two orders of magnitude, observed 100-fold difference in the promoter activities.
while only a two- to threefold variation was observed in the Among the 10 studied dehaloboxes, only DB7 is a perfect
yield of functional (i.e. soluble) production of CprK inverted repeat (TTAAT-N4-ATTAA). In addition, it is
homologues by E. coli JM109(DE3) (Fig. 1c). Hence, the notable that among all the promoters, that of cprBA1

Fig. 3. In vivo promoter probe assays using recombinant E. coli JM109(DE3) cells overproducing CprK2 (a), CprK1 (b) or
CprK4 (c). Dehaloboxes (DB1–10) centred in D. hafniense promoter fragments were fused to b-galactosidase-encoding genes
and tested for transcriptional activation in the presence of 20 mM CHPA (black bars) or in the absence of the effector (grey
bars). Results shown in (b) were obtained previously (Gábor et al., 2006) and are included for comparison. Error bars, SDs for
mean values from three biological replicates.

http://mic.sgmjournals.org 3691
K. Gábor and others

containing DB7 shows the highest promoter activity, not be synthesized in D. hafniense); or alternatively, the effector-
only in the presence of CprK1 (500 Miller units; Fig. 3b) free CprK4 is already capable of DNA binding and this
but also with CprK4 (130 Miller units; results not shown). ability is only enhanced in the presence of a true ligand
This makes DB7 a candidate optimal binding target for molecule (apparently not by CHPA). In order to study the
CprK-like proteins. We investigated the correlation DNA-binding properties of CprK4 in the presence of
between the promoter activity and changes from the different potential effector molecules, the protein was
optimal target sequence (DB7, gTTAATacacATTAAt). purified and EMSAs were performed. This in vitro method
First, for each nucleotide position of the dehaloboxes, a overcomes the difficulties raised by the toxic nature of these
weight measure (conservation value) was assigned using compounds when used in growth experiments. In parallel,
the Weblogo 3 BETA program. A high value indicated a the alternative effector range of CprK1 was mapped with
highly conserved position, while a low number suggested the same method. Attempts to functionally purify CprK2
that there is no preferred nucleotide at a certain position. were unsuccessful, due to the instability of the protein;
Next, we compared each dehalobox to DB7. At the therefore, CprK2 was excluded from the EMSA experi-
positions that showed discrepancies from the optimal ments. Binding experiments were performed with small
binding sequence, the sums of the weight measure values promoter fragments containing dehaloboxes for which the
were taken as an indication of the degree of discrepancy. highest in vivo activity was measured previously (DB7 for
Finally, we computed the correlation between the degree of CprK1 and DB8 for CprK4; Fig. 3). Labelled DNA
discrepancy and the corresponding promoter activity. fragments (1.5–2 nM) were incubated in the presence of
Correlation was measured on a 0–1 scale, where values 1 mM purified regulator protein and 800 mM potential
up to 0.4 show negligible to weak correlation, while values effector molecules. These chemicals included cAMP, 4-
above 0.7 indicate a strong to very strong correlation hydroxyphenylacetic acid (HPA) (a dechlorinated deriv-
between two variables. A strong negative correlation ative of CHPA), di- and trichlorinated phenolic compounds
(20.76) was observed between the measured reporter that have been reported to be degraded by D. hafniense
enzyme activity and the degree of discrepancy from the DCB-2 and/or the closely related D. dehalogenans [3,5-DCP,
putative optimal target sequence TTAAT-N4-ATTAA (Fig. 2,3-DCP, 2,4-DCP, 2,4,5-trichlorophenol (TCP) and 2,4,6-
4). Similar results were obtained (correlation coefficient TCP] (Christiansen & Ahring, 1996; Utkin et al., 1995), and
was 20.68) when changes in the consensus Pribnow CHPA and its bromo- and fluoro-substituted derivatives
sequence (TATAAT) were also taken into account. (2-Br-4-CP and 2-F-4-CP). The experiments revealed that
CprK4 is indeed capable of DNA binding in its effector-free
form (Fig. 5b), and that cAMP or CHPA do not promote
Alternative effectors of CprK1 and CprK4
DNA binding to any extent. However, in the presence of
In vivo promoter probe experiments in E. coli revealed an compounds such as 3,5-DCP, 2,3-DCP and 2,4,6-TCP, over
unusual feature of CprK4: although it distinguished 80 % of the DNA formed a complex with CprK4. It is
between preferred (DB8) and non-preferred (DB9 and interesting to note that exactly these compounds had no
DB10) D. hafniense promoter fragments, it was constitu- potential to act as effectors for CprK1 (Fig. 5a). This
tively active, irrespective of the presence or absence of the suggests a complementary function of the two paralogues.
potential effector CHPA (Fig. 3c). There are two possible The analogues 2,4-DCP and CHPA are very good effectors
explanations for this behaviour: either CprK4 was activated for CprK1, similarly to 2-Br-4-CP. The related compound
by cAMP (a molecule that is present in E. coli but cannot that contained a fluorine atom (2-F-4-CP) was a weak
effector, similarly to 2,4,5-TCP. Unexpectedly, in the
presence of a high molar excess of HPA, protein–DNA
complexes were also formed. In the case of CprK4, addition
of 2,4-DCP, 2-Br/F-4-CP and 2,4,5-TCP resulted in an
intermediate level of increase in protein–DNA complex
formation (70–75 % complexed from the total DNA). In
summary, the results indicate that a halogen substitution at
the meta position in phenol derivatives resulted in the loss
of effector activity on CprK1; in contrast, this appeared to
be beneficial for the activity of CprK4 (Fig. 5c). The only
restriction for CprK4 is that it cannot accept a long side-
chain (such as the acetic acid tail in CHPA) at the para
Fig. 4. Correlation between promoter activity mediated by the position, probably due to steric constraints in the effector-
CprK paralogues and the degree of discrepancy from the optimal binding pocket. These results are in agreement with those
regulator-binding sequence. Promoter activity is expressed in obtained by micro electrospray ionization MS (mESI-MS)
Miller units (enzymic activity of the reporter protein); the degree of
(Mazon et al., 2007).
discrepancy is calculated as the sum of the weight measure values
of each nucleotide position that differs from the optimal dehalobox The last notable difference between the in vitro DNA-
sequence DB7 (gTTAATacacATTAAt). binding properties of the two proteins is the formation of a

3692 Microbiology 154

 Transcriptional activation by CprK paralogues

Fig. 5. Alternative effectors of CprK1 and CprK4. EMSA was performed in the presence of halogenated compounds and
compounds with related structure (as indicated below the image), and with constant amounts of CprK1 (a) or CprK4 (b). The
reaction mixture contained dehalobox DB7 (CprK1) or DB8 (CprK4) and 800 : 1 molar ratios of effector and protein. The first
lane on the left in each gel served as a control, containing only labelled DNA. Free DNA is indicated by open arrowheads,
protein–DNA complexes are indicated by filled arrowheads. Free DNA or DNA in complex with the regulator was quantified as a
percentage of the total DNA loaded in each well. (c) The positions in phenol derivatives where halogen or other substitutions
can or cannot allow DNA–protein complex formation by CprK1 or CprK4.

second, slower-migrating protein–DNA complex with DISCUSSION

CprK4, which is not observed with CprK1 (Fig. 5b). This
indicates that more CprK4 molecules are bound to the Whole-genome shotgun sequencing of the D. hafniense
same DNA fragment, causing a further decrease in its DCB-2 genome revealed a redundancy in genes encoding
mobility, either by forming a higher CprK4 oligomer (most potential reductive dehalogenases (cprA) and putative
likely a tetramer, as proposed by Joyce et al., 2006), which CRP–FNR-type transcriptional regulators (cprK), which
then binds to the DNA, or by the binding of a second enables this organism to gain energy from halorespiration
CprK4 dimer to a weak, low-affinity DNA site. One weak (Smidt & de Vos, 2004; Villemur et al., 2002). Moreover,
inverted repeat (TTATC-N4-ATTTA) that resembles the these cprA- and cprK-like genes were found to be located in
dehalobox consensus was indeed identified at a position a conserved gene context resembling those found in other
that overlaps DB8. The possible, although fairly unlikely, halorespiring bacteria (Smidt et al., 2000). The work
functionality of this putative binding site could give an presented here focused on the diversity of predicted CprK
explanation of how CprK4-mediated transcriptional activa- regulators and aimed to gain more understanding of
tion can be controlled by a co-operative action of a second the roles of multiple CprK homologues in strain DCB-2
transcriptional regulator. (overlapping or complementary function). In this

http://mic.sgmjournals.org 3693
K. Gábor and others

discussion we also suggest how their specificity is exerted, activity, it is possible that in D. hafniense the expression of
or in other words, how cross-talk of these highly similar cprK4 is regulated in a hierarchical way, as described for
regulators is eliminated, and what is the origin of the FixK in Rhizobium meliloti (Batut et al., 1989). The
redundancy of the CprK-like proteins. expression of fixK is activated in an oxygen-dependent
manner by the two-component signal transduction pro-
We found that the CprK paralogues in D. hafniense DCB-2
teins FixLJ. Also, hierarchy can be exerted by means of
have at least partly overlapping functions, as both CprK1
other regulatory interactions. As an example, in Paracoccus
and CprK2 initiated transcription from dehalobox-con-
denitrificans the role of three FNR homologues (FnrP,
taining promoters in the presence of CHPA. CprK1 and
NNR and NarR) is separated by the need for the presence
CprK2 are 62 % similar at the amino acid level, which
might indicate that the two regulators can substitute for of a specific sigma factor (Van Spanning et al., 1997;
each other, and that they both control regulons with Veldman et al., 2006). However, in the CprK paralogues
identical, or at least overlapping, functionality. Such most of the characteristic negatively charged residues that
interchangeable function has been demonstrated in make contact with the common s70 factor of the D.
Lactococcus lactis, in which two closely related paralogues hafniense RNAP holoenzyme are conserved; thus, it is
from the CRP–FNR family are present: FlpA and FlpB unlikely that they utilize an alternative sigma factor (Gábor
(75 % similarity), encoded by distinct operons involved in et al., 2006). The presence of co-activators or repressors
metal-ion transport. The single L. lactis Flp mutants behave can also alter promoter activity (Barnard et al., 2004;
similarly to the wild-type cells; only the double mutant Bearson et al., 2002). The results from EMSA experiments
shows an altered polypeptide profile and drastically indicate that the promoter regulated by CprK4 might
decreased zinc content (Gostick et al., 1999). We also contain one additional, weak-affinity regulator-binding
found that one CprK paralogue (CprK4) responds to a site, as shown by the formation of a second, slower-
different set of effector molecules, indicating that it has a migrating band (Fig. 5b). This distinct complex might have
function complementary to that of CprK1. Specific DNA deviating features, e.g. serving as the target for a co-
binding mediated by CprK4 was enhanced in the presence regulator. The third possibility is that the physical shape of
of meta-substituted compounds such as 3,5-DCP and 2,3- the DNA in complex with CprK4 is different under
DCP. In contrast, these compounds did not induce DNA effector-free and effector-bound conditions, as has been
binding by CprK1. The recently elucidated crystal structure shown for LysR-type regulators (Diaz & Prieto, 2000),
of CprK1 reveals that such differences in effector specificity which would enable RNAP recruitment only under
are expected, since among the CHPA-binding residues of halorespiring conditions in D. hafniense.
CprK1, only Lys133 is conserved in all the CprK paralogues Finally, we have demonstrated that at least three CprK
(CprK4 shares five identical effector-binding amino acids paralogues are involved in sensing a range of halogenated
with CprK1) (Fig. 2) (Joyce et al., 2006). In conclusion, we compounds in D. hafniense. Genomic analyses have
showed that CprK2 and CprK4 can sense a range of revealed that a large number of gene paralogues may be
halogenated phenol derivatives in a co-operative manner, present in bacterial chromosomes. It is known that gene
with only partly overlapping specificity. duplications and horizontal gene transfer (HGT), which
The large number of transcriptional regulators from the give rise to gene paralogues, play an important role in the
CRP–FNR family in the same organism raises the question evolution of bacterial genomes by providing a rapid means
of how target genes are discriminated in the simultaneous of adaptation to environmental changes (Hurles, 2004;
presence of these regulators. It has been speculated that Janssen et al., 2005; Van der Meer & Sentchilo, 2003). An
subtle changes in the target DNA sequence, the exertion of interesting theory of Parales & Harwood (2002) suggests
hierarchical control and further protein requirements (co- that chemotaxis provides an excellent opportunity to
activators and specific sigma factors) ensure the elimina- promote HGT. It directs motile bacteria towards con-
tion of unwanted cross-talk of the similar regulators taminated sites where strains carrying the relevant catabolic
(Zumft, 1997). We showed that a strong negative plasmids are likely to be present and therefore can donate
correlation exists between deviations from the optimal new catabolic genes to the recipient chemotactic bacteria.
CprK-binding sequence (TTAAT-N4-ATTAA) and the Remarkably, our results indicate that the promoter of
promoter activity determined by using lacZ promoter macA, a putative methyl-accepting chemotaxis protein-
fusions in E. coli. Conservation of the A/T and T/A base encoding gene, is recognized by CprK4. D. hafniense DCB-
pairs at the symmetry-related positions 3 and 12 in 2 possesses one or two terminal flagella (Christiansen &
dehaloboxes ensures that these promoters are not recog- Ahring, 1996), and with one exception (the phosphopro-
nized by proteins with the FNR-type E–SR motif in their tein phosphatase cheZ) all the che genes required for the
recognition a-helix (Gábor et al., 2006). It has also been chemotaxis machinery are present in the genome: the core
demonstrated that both the spacing nucleotides of the receptor-encoding cheA, cheY and cheW, and the modu-
inverted repeat and the nucleotides further downstream lating cheR and cheB (Webre et al., 2004); therefore, it is
affect transcriptional activation, which enables fine tuning tempting to assume that this might enable the bacterium to
of promoter activity (Scott et al., 2003; Veldman et al., detect and respond to specific chemicals (potentially
2006). As CprK4 showed a constitutive DNA-binding including organohalides) in the environment. Hence, it is

3694 Microbiology 154

 Transcriptional activation by CprK paralogues

possible that this particular CprK homologue regulates not Christiansen, N., Ahring, B. K., Wohlfarth, G. & Diekert, G. (1998).
only genes that are involved in the degradation of Purification and characterization of the 3-chloro-4-hydroxyphenyla-
cetate reductive dehalogenase of Desulfitobacterium hafniense. FEBS
haloaromatic compounds but also a gene that enables the
Lett 436, 159–162.
organism to actively seek these compounds. Such a
Diaz, E. & Prieto, M. A. (2000). Bacterial promoters triggering
relationship has been described in the 4-hydroxybenzoate
biodegradation of aromatic pollutants. Curr Opin Biotechnol 11, 467–
(4-HB)-degrading Pseudomonas putida PRS2000 strain 475.
between the PcaR transcriptional regulator and PcaK,
Egland, P. G. & Harwood, C. S. (2000). HbaR, a 4-hydroxybenzoate
which is involved in the chemotaxtic response to 4-HB sensor and FNR–CRP superfamily member, regulates anaerobic 4-
(Harwood et al., 1994), and for NahR and NahY from P. hydroxybenzoate degradation by Rhodopseudomonas palustris. J
putida G7 for the degradation of, and chemotaxis to the Bacteriol 182, 100–106.
aromatic hydrocarbon naphthalene (Grimm & Harwood, Gábor, K., Verı́ssimo, C. S., Cyran, B. C., Ter Horst, P., Meijer, N. P.,
1999; Schell, 1985). Smidt, H., de Vos, W. M. & van der Oost, J. (2006). Characterisation
of CprK1, a CRP/FNR-type transcriptional regulator of halorespira-
In conclusion, we have investigated the role of multiple tion from Desulfitobacterium hafniense. J Bacteriol 188, 2604–2613.
CprK paralogues in D. hafniense DCB-2, and found that at
Gostick, D. O., Griffin, H. G., Shearman, C. A., Scott, C., Green, J.,
least two regulators (CprK2 and CprK4) have a distinct
Gasson, M. J. & Guest, J. R. (1999). Two operons that encode FNR-
effector-sensing function. The DNA-binding specificity is like proteins in Lactococcus lactis. Mol Microbiol 31, 1523–1535.
possibly exerted by changes in the target nucleotide
Green, J., Scott, C. & Guest, J. R. (2001). Functional versatility in the
sequence from the consensus dehalobox sequence. The CRP–FNR superfamily of transcription factors: FNR and FLP. Adv
redundancy of the cprK genes is likely to be caused by Microb Physiol 44, 1–34.
mobile genetic elements, as shown by the presence of a Grimm, A. C. & Harwood, C. S. (1999). NahY, a catabolic plasmid-
putative transposase-encoding gene in the vicinity of the encoded receptor required for chemotaxis of Pseudomonas putida to
cpr gene clusters. The presence of a halorespiration- the aromatic hydrocarbon naphthalene. J Bacteriol 181, 3310–3316.
inducible methyl-accepting chemotaxis protein-encoding Harwood, C. S., Nichols, N. N., Kim, M. K., Ditty, J. L. & Parales, R. E.
gene in D. hafniense suggests that chemotaxis plays an (1994). Identification of the pcaRKF gene cluster from Pseudomonas
important role in actively seeking halogenated compounds. putida: involvement in chemotaxis, biodegradation, and transport of
4-hydroxybenzoate. J Bacteriol 176, 6479–6488.
Hurles, M. (2004). Gene duplication: the genomic trade in spare parts.
ACKNOWLEDGEMENTS PLoS Biol 2, E206.
The sequence data for the D. hafniense DCB-2 genome were produced Israelsen, H., Madsen, S. M., Vrang, A., Hansen, E. B. & Johansen, E.
by the US Department of Energy Joint Genome Institute (http:// (1995). Cloning and partial characterization of regulated promoters
www.jgi.doe.gov). We thank Eric Johansen (Chr. Hansen, Denmark) from Lactococcus lactis Tn917-lacZ integrants with the new promoter
for kindly providing the pAK80 plasmid. probe vector, pAK80. Appl Environ Microbiol 61, 2540–2547.
Janssen, D. B., Dinkla, I. J., Poelarends, G. J. & Terpstra, P. (2005).
Bacterial degradation of xenobiotic compounds: evolution and
REFERENCES distribution of novel enzyme activities. Environ Microbiol 7, 1868–1882.
Joyce, M. G., Levy, C., Gábor, K., Pop, S. M., Biehl, B. D., Doukov, T. I.,
Anjum, M. F., Green, J. & Guest, J. R. (2000). YeiL, the third member Ryter, J. M., Mazon, H., Smidt, H. & other authors (2006). CprK
of the CRP–FNR family in Escherichia coli. Microbiology 146, 3157– crystal structures reveal mechanism for transcriptional control of
3170. halorespiration. J Biol Chem 281, 28318–28325.
Bailey, T. L. & Elkan, C. (1994). Fitting a mixture model by Kolb, A., Busby, S., Buc, H., Garges, S. & Adhya, S. (1993).
expectation maximization to discover motifs in biopolymers. Proc Int Transcriptional regulation by cAMP and its receptor protein. Annu
Conf Intell Syst Mol Biol 2, 28–36. Rev Biochem 62, 749–795.
Barnard, A., Wolfe, A. & Busby, S. (2004). Regulation at complex Korner, H., Sofia, H. J. & Zumft, W. G. (2003). Phylogeny of the
bacterial promoters: how bacteria use different promoter organiza- bacterial superfamily of Crp–Fnr transcription regulators: exploiting
tions to produce different regulatory outcomes. Curr Opin Microbiol the metabolic spectrum by controlling alternative gene programs.
7, 102–108. FEMS Microbiol Rev 27, 559–592.
Batut, J., Daveran-Mingot, M. L., David, M., Jacobs, J., Garnerone, Madsen, T. & Licht, D. (1992). Isolation and characterization of an
A. M. & Kahn, D. (1989). fixK, a gene homologous with fnr and crp anaerobic chlorophenol-transforming bacterium. Appl Environ
from Escherichia coli, regulates nitrogen fixation genes both positively Microbiol 58, 2874–2878.
and negatively in Rhizobium meliloti. EMBO J 8, 1279–1286. Mazon, H., Gábor, K., Leys, D., Heck, A. J., van der Oost, J. & van den
Bauer, C. E., Elsen, S. & Bird, T. H. (1999). Mechanisms for redox Heuvel, R. H. (2007). Transcriptional activation by CprK1 is regulated
control of gene expression. Annu Rev Microbiol 53, 495–523. by protein structural changes induced by effector binding and redox
Bearson, S. M., Albrecht, J. A. & Gunsalus, R. P. (2002). Oxygen and state. J Biol Chem 282, 11281–11290.
nitrate-dependent regulation of dmsABC operon expression in McGinnis, S. & Madden, T. L. (2004). BLAST: at the core of a powerful
Escherichia coli: sites for Fnr and NarL protein interactions. BMC and diverse set of sequence analysis tools. Nucleic Acids Res 32, W20–
Microbiol 2, 13. W25.
Christiansen, N. & Ahring, B. K. (1996). Desulfitobacterium hafniense Mesa, S., Hennecke, H. & Fischer, H. M. (2006). A multitude of CRP/
sp. nov., an anaerobic, reductively dechlorinating bacterium. Int J Syst FNR-like transcription proteins in Bradyrhizobium japonicum.
Bacteriol 46, 442–448. Biochem Soc Trans 34, 156–159.

http://mic.sgmjournals.org 3695
K. Gábor and others

Nicholas, K. B., Nicholas, H. B., Jr & Deerfield, D. W., II (1997). Van der Meer, J. R. & Sentchilo, V. (2003). Genomic islands and the
GeneDoc: analysis and visualization of genetic variation. EMBnet.news evolution of catabolic pathways in bacteria. Curr Opin Biotechnol 14,
4 (2), 1–4. 248–254.
Parales, R. E. & Harwood, C. S. (2002). Bacterial chemotaxis to Van Eekert, M. H. & Schraa, G. (2001). The potential of anaerobic
pollutants and plant-derived aromatic molecules. Curr Opin Microbiol bacteria to degrade chlorinated compounds. Water Sci Technol 44,
5, 266–273. 49–56.
Pop, S. M., Kolarik, R. J. & Ragsdale, S. W. (2004). Regulation of Van Spanning, R. J., De Boer, A. P., Reijnders, W. N., Westerhoff,
anaerobic dehalorespiration by the transcriptional activator CprK. J H. V., Stouthamer, A. H. & Van der Oost, J. (1997). FnrP and NNR of
Biol Chem 279, 49910–49918. Paracoccus denitrificans are both members of the FNR family of
Schell, M. A. (1985). Transcriptional control of the nah and sal transcriptional activators but have distinct roles in respiratory
hydrocarbon-degradation operons by the nahR gene product. Gene adaptation in response to oxygen limitation. Mol Microbiol 23, 893–
36, 301–309. 907.
Schneider, T. D. & Stephens, R. M. (1990). Sequence Logos: a new Veldman, R., Reijnders, W. N. & Van Spanning, R. J. (2006).
way to display consensus sequences. Nucleic Acids Res 18, 6097–6100. Specificity of FNR-type regulators in Paracoccus denitrificans.
Scott, C., Partridge, J. D., Stephenson, J. R. & Green, J. (2003). DNA Biochem Soc Trans 34, 94–96.
target sequence and FNR-dependent gene expression. FEBS Lett 541, Villemur, R., Saucier, M., Gauthier, A. & Beaudet, R. (2002).
97–101. Occurrence of several genes encoding putative reductive dehalo-
Smidt, H. & de Vos, W. M. (2004). Anaerobic microbial dehalogena- genases in Desulfitobacterium hafniense/frappieri and Dehalococcoides
tion. Annu Rev Microbiol 58, 43–73. ethenogenes. Can J Microbiol 48, 697–706.
Smidt, H., van Leest, M., van der Oost, J. & de Vos, W. M. (2000). Vollack, K. U., Hartig, E., Korner, H. & Zumft, W. G. (1999). Multiple
Transcriptional regulation of the cpr gene cluster in ortho-chlor- transcription factors of the FNR family in denitrifying Pseudomonas
ophenol-respiring Desulfitobacterium dehalogenans. J Bacteriol 182, stutzeri: characterization of four fnr-like genes, regulatory responses
5683–5691. and cognate metabolic processes. Mol Microbiol 31, 1681–1694.
Thompson, J. D., Gibson, T. J., Plewniak, F., Jeanmougin, F. & Webre, D. J., Wolanin, P. M. & Stock, J. B. (2004). Modulated receptor
Higgins, D. G. (1997). The CLUSTAL_X windows interface: flexible interactions in bacterial transmembrane signaling. Trends Cell Biol 14,
strategies for multiple sequence alignment aided by quality analysis 478–482.
tools. Nucleic Acids Res 25, 4876–4882. Zumft, W. G. (1997). Cell biology and molecular basis of denitrifica-
Utkin, I., Dalton, D. D. & Wiegel, J. (1995). Specificity of reductive tion. Microbiol Mol Biol Rev 61, 533–616.
dehalogenation of substituted ortho-chlorophenols by Desulfitobacter-
ium dehalogenans JW/IU-DC1. Appl Environ Microbiol 61, 346–351. Edited by: J. Green

3696 Microbiology 154