Regular paper on-line at: www.actabp.pl In vitro DNA binding of purified CcpA protein from

During this study His-tagged CcpA protein purified under native conditions to obtain a biologically active protein was used for molecular analysis of CcpA-dependent regulation. Using electrophoretic mobility shift assays it was demonstrated that CcpA of L. lactis can bind DNA in the absence of the HPr-Ser-P corepressor and exhibits DNA-binding affinity for nucleotide sequences lacking cre sites. However, purified HPr-Ser-P protein from Bacillus subtilis was shown to slightly increase the DNA-binding capacity of the CcpA protein. It was also observed that CcpA bound to the cre box forms an apparently more stable complex than that resulting from unspecific binding. Competition gel retardation assay performed on DNA sequences from two PEP:PTS regions demonstrated that the ybhE, bglS, rheB, yebE, ptcB and yecA genes situated in these regions are most probably directly regulated by CcpA.


INTRODUCTION
In Gram-positive bacteria carbon catabolite repression is exerted through binding of the CcpA protein to cis-acting catabolite responsive element (cre).The cre site was first identified in Bacillus subtilis (Nicholson et al., 1987;Miwa & Fujita, 1990).The 14-nucleotide consensus sequence TGWNANCGNT-NWCA (W, adenine or thymine) was proposed based on site-directed mutagenesis studies (Weickert & Chambliss, 1990).CcpA interacts with the HPr-Ser-P protein, which distinguishes this regulator from other regulators of the LacI/GalR family that interact with low molecular mass compounds (Schumacher et al., 2004).In B. subtilis, the complex of CcpA and seryl-phosphorylated HPr protein can bind to cre elements located in the vicinity of regu lated genes.The physiological effect depends on the position of the cre site in respect to the regulated gene (Titgemeyer & Hillen, 2002).CcpA acts as an activator when the cre sequence is located upstream of the promoter sequence of the regulated gene.On the other hand, binding of this regulator to cre located either in the promoter region of catabolic genes or situated within the coding sequence prevents, respectively, the initiation of transcription or transcript elongation.
Crystal structures for the B. megaterium apoC-cpA protein and the CcpA-(HPr-Ser-P)-DNA complex were reported recently (Schumacher et al., 2004).These results showed that CcpA has a three-dimensional structure similar to the regulators of the LacI/ GalR family and confirmed the presence of a helixturn-helix (HTH) DNA-binding N-terminal domain, an effector binding site and a C-terminal dimerisation domain (Henkin et al., 1991;Weickert & Adhya, 1992;Schumacher et al., 2004).
The Lactococcus lactis subsp.lactis ccpA gene (Gene Bank accession No. AF106673) was identified and sequenced in our group (Aleksandrzak et al., 2000).The 37 kDa gene product shares 50% identity with CcpA from B. subtilis (Henkin et al., 1991; UniProtKB/Swiss-Prot P25144), while the identity to CcpA from L. lactis ssp.cremoris (Luesink et al., 1998; UniProtKB/TrEMBL Q9ZAE0) reaches 96%.Most data on the CcpA proteins and their interactions come from studies of CcpA in the Bacillus species.To our knowledge, there are no data describing interaction of L. lactis CcpA with DNA.
In this study, we used His-tagged CcpA protein from L. lactis subsp.lactis IL1403, which was overproduced in Escherichia coli, purified under native conditions (Kowalczyk & Bardowski, 2003), and subsequently used for molecular analysis of its interactions with DNA.
To analyse the DNA-binding ability of the CcpA protein from L. lactis we used electrophoretic mobility shift assays (EMSA).Our results showed that CcpA of L. lactis can bind DNA in the absence of the HPr-Ser-P corepressor and is able to bind DNA lacking cre sites.However, it seems that CcpA has a higher affinity for cre-containing fragments and that HPr-Ser-P slightly increases the DNA-binding ability of the CcpA protein.
In the L. lactis IL1403 genome several thousand cre-like sequences can be detected, yet most of them have no biological significance because of their unsuitable location (Guédon et al., 2002).We found that CcpA is involved in the regulation of several genes coding for proteins engaged in cellobiose and lactose transport or catabolism in L. lactis (Kowalczyk et al., unpublished).These genes are localised in two different regions within the L. lactis genome, which for the purpose of this publication were termed "PEP:PTS regions".A search for the L. lactis cre TGNNANCGNTNNCA motif proposed by E. Guédon and coworkers allowing one or two mismatches except for the central CG base pairs (Guédon et al., 2002), showed several cre-like sites in the PEP:PTS regions.To elucidate whether the CcpA effect is direct or indirect, binding of CcpA to crecontaining DNA fragments located within two PEP: PTS regions was studied (Fig. 1).The obtained results demonstrated that some of the analysed genes might be directly regulated by the CcpA protein.

MATERIALS AND METHODS
Recombinant DNA and PCR methods.General recombinant DNA methods were as described previously (Sambrook et al., 1989).DNA fragments were amplified by PCR with 1 U of Taq polymerase (Polgen) using appropriate primers presented in Table 1.
The PCR reaction was performed for 30 cycles at 95 o C for 1 min, 55 o C for 30 s and 72 o C for 30 s. PCR products were purified with Montage PCR Centrifugal Filter Devices (Millipore).DNA was quantified by gel electrophoresis with Smart Ladder (Eurogentec) and also spectrophotometrically at 260 nm (1 A 260 = 50 µg/ml).
The oligonucleotides for electrophoretic mobility shift assay were obtained from the DNA Sequencing and Oligonucleotide Synthesis Lab of our Institute.To produce double-stranded DNA fragments, complementary oligonucleotides at 100 ng each were mixed and incubated at 95 o C for 3 min in a water bath and then cooled down slowly to the room temperature.
For the DNA binding in the presence of HPr-Ser-P and competition experiments, DNA was 5'- Sequences are derived from the L. lactis IL1403 genome (AE005176).The potential cre elements deviating from the cre consensus TGNNANCGNTNNCA by a maximum of two mismatches, but not in the central CG base pairs, were identified by sequence analysis.The cre elements that were further analysed in this study are marked as vertical black lines whereas other cre sites are marked as vertical grey lines.Genes encoding proteins homologous to the A, B and C components of the PEP:PTS system are denoted in black and those homologous to hypothetical regulators in grey.Numbering of cre sites given here refers directly to the numbering of lanes in Fig. 7. Point marked as "0" represents ybhE cre.labelled using T 4 polynucleotide kinase and [γ-32 P] ATP (Fermentas), and purified with the QIAquick PCR purification kit (Qiagen).Binding was carried out in 20 µl of buffer A with 25 mM d-fructose-1,6bisphosphate and 2 µg/ml of sonicated calf thymus DNA to reduce non-specific binding.In competition experiments, after 10 min of incubation at 30°C unlabelled competing DNA was added and incubated for another 10 min.Samples were mixed with 2 µl of 50% glycerol containing 0.1% bromophenol blue and loaded onto a 5% polyacrylamide gel.
DNA electrophoresis.For general purposes, DNA electrophoresis was performed in TAE buffer on agarose gels supplemented with ethidium bromide.The gel concentration and agarose type varied according to the sizes of separated DNA fragments (Sambrook et al., 1989).
For EMSA at 4°C, 1.5% MetaPhor agarose supplemented with ethidium bromide in TBE buffer was used.In the case of EMSA with oligonucleotides the concentration of MetaPhor agarose was increased to 3%.After electrophoresis, gels were photographed under UV light for documentation.
For DNA binding in the presence of HPr-Ser-P and competition experiments the EMSA in 5% polyacrylamide (acrylamide/bis-acrylamide, 82:1) gel was performed.Horizontal gels were pre-run in 0.5 × TBE buffer for 30 min at 3 V/cm.After electrophoresis in the same buffer for 60 to 90 min at 10 V/cm, the gel was dried onto filter paper, and radiolabelled bands were visualised by autoradiography.

Model structure of CcpA in complex with both DNA and HPr
As the identity of B. megaterium CcpA and L. lactis CcpA sequences is lower than 60%, the already determined three-dimensional structures of B. megaterium CcpA (apo-form and in complex with HPr and DNA) (Schumacher et al., 2004) cannot be directly extrapolated onto the lactococcal CcpA spatial structure.Yet, the crystallographic data obtained for B. megaterium CcpA were used as a basis to create a hypothetical model of the three-dimensional structure of the CcpA protein of L. lactis.
Atomic coordinates of the L. lactis CcpA 3D model were obtained using the SWISS-MODEL server (Peitsch, 1995;Guex & Peitsch, 1997;Schwede et al., 2003).A model structure of the L. lactis CcpA dimer interacting with DNA and HPr was created by superposition of two modelled lactococcal CcpA monomers on the B. megaterium CcpA dimer (Fig. 2).
According to the model, the N-terminal domains of the CcpA dimer interact with DNA.Fragments I8-N27 (HTH motif) and A53-S60 (hinge helix) are proposed to directly contact DNA.As the first six N-terminal residues are located outside the protein-DNA interface, the His-tag at the N-terminus should not disturb the CcpA-DNA interactions (Fig. 2B).
These suppositions were recently confirmed by a collaborative study between our group and the group of W. Sanger through analysis of the L. lactis CcpA crystal structure (Loll et al., in press).This analysis showed that CcpA residues which are involved in the interaction with DNA or with the HPr protein as well as the overall structure of the protein are conserved between bacilli and lactococci.

Binding of lactococcal CcpA to DNA
It has previously been shown that catabolite repression depends on the CcpA protein and a cis-

Name
Oligonucleotide* ybhEcref ybhEcrer 5'-ATCAGGCGATTCTTCGTGAATA-3' 5'-TGAGCGTTCTGGATAGATAGAT-3' nocref2 nocrer2 5'-ACCTCCAATTGCTCCAGAAA-3' 5'-GTCAAGGCCTTAAACTC-3' ocreybhf ocreybhr 5'-TTTTAAGATGATATCGCTTGCAAAAATAAA-3' 5'-TTTATTTTTGCAAGCGATATCATCTTAAAA-3' et al., 2001.acting catabolite responsive element (cre) located near the promoter region of the regulated gene (Nicholson et al., 1987;Weickert & Chambliss, 1990;Henkin et al., 1991).CcpA has a helix-turn-helix motif responsible for DNA binding and its binding to DNA in B. subtilis, B. megaterium and lactobacilli was demonstrated by several authors (Miwa et al., 1994;Kim et al., 1995;Jones et al., 1997;Schick et al., 1999;Aung-Hilbrich et al., 2002;Mahr et al., 2002).In this work, the ability of the L. lactis CcpA protein to bind DNA was analysed using EMSA (Fig. 3).For this analysis two PCR-generated DNA fragments were synthe-sised -one of 400 bp (AE005176, 175977..176376 bp) containing a cre site consistent with the cre consensus (ybhE cre) and the second one of 316 bp (AE005176, 2107756..2108071 bp) without cre (no-cre).Both DNA fragments were separately incubated in buffer A with 2 μg of CcpA or 2 μg of BSA (bovine serum albumin) as a control protein.After electrophoresis of the reaction mixtures both fragments -without cre (Fig. 3; lanes 4, 5) and with the ybhE cre (Fig. 3; lanes 8, 9) were shifted by the CcpA protein.However, much lower amounts of DNA were retarded in the case of the no-cre DNA fragment, for which a smear was visible below the DNA/CcpA complex.The use of BSA in place of CcpA did not lead to formation of any detectable complex (Fig. 3; lanes 3, 7) nor did it significantly change the formation of the DNA/CcpA complex when added at the same time as CcpA (Fig. 3; lanes 5, 9).Since retardation of DNA fragments was observed for both no-cre and ybhE cre, a similar experiment was performed in buffer B used by Aung-Hilbrich et al. (2002).In this buffer, discrimination between specific and unspecific binding was expected to be observed.Both DNA fragments, nocre and the ybhE cre sequence, were retarded in the presence of CcpA (not shown).However, the amount of DNA shifted was lower than that detected in buffer A.

DNA binding of CcpA from Lactococcus lactis
Hence, to minimise the unspecific binding of CcpA to DNA, short oligonucleotides were used (Fig. 4).Double-stranded oligonucleotides: onocre, containing no cre (Fig. 4, lanes 1-3), okcre, the sequence used by Aung-Hilbrich and coworkers (Fig. 4, lanes 4-6), and ocreybhE containing the cre sequence from the promoter region of the ybhE gene (Fig. 4, lanes 7-9) were used.All the oligonucleotides tested were shifted (Fig. 4, lanes 3, 6, 9), suggesting that the unspecific binding of CcpA to DNA does not depend on the length of DNA.

Effect of HPr-Ser-P on CcpA binding to DNA
The phosphorylated HPr corepressor protein was shown to be necessary for catabolite repression and formation of a specific ternary complex with cre-containing DNA in B. megaterium (Jones et al., 1997;Schumacher et al., 2004).Therefore, in order to increase the specific binding of lactococcal CcpA to DNA containing cre sites the HPr-Ser-P protein of B. subtilis was used.For binding, the concentrations of the CcpA protein that enabled observing partial shifting of DNA and an excess of HPr-Ser-P which binds to the surface of CcpA at the stoichiometry of 2 molecules per 1 CcpA dimer were used (Fig. 5).
In a control experiment it was observed that the addition of HPr-Ser-P to the reaction mixture containing CcpA and the DNA fragment without cre sequence did not change the efficiency of CcpA binding to DNA (not shown).In contrast, addition of HPr-Ser-P from B. subtilis to reaction mixtures containing CcpA and the DNA fragment with cre sequence slightly increased the CcpA binding (Fig. 5).

Competition EMSA with cre-containing DNA fragments located within two PEP:PTS regions
As the specific and unspecific DNA binding of the CcpA protein was difficult to discriminate, the affinity of CcpA for various cre sites and to nocre fragments was examined by competition EMSA.First, unlabelled ybhE cre site (Fig. 6, lanes 3, 5, 7, 9) and a no-cre fragment (Fig. 6, lanes 4, 6, 8, 10) were tested for the ability to sequester CcpA from the labelled ybhE cre probe.The CcpA/ybhE cre complexes (Fig. 6, lanes 4, 6, 8, 10) were observed to migrate very similarly in all samples irrespective of the presence of the non-labelled no-cre fragment.In contrast, the labelled ybhE cre fragments (Fig. 6, lanes 3, 5, 7, 9) migrated faster, as more dispersed bands than the control (Fig. 6, lane 2) when challenged with the non-labelled ybhE cre competitor.Altogether, these results demonstrated that the unlabelled ybhE cre fragment, unlike no-cre, was able to replace labelled ybhE cre from the CcpA/cre complex.
Competition EMSA was also performed with several other unlabelled cre-containing fragments located within two PEP:PTS regions (Figs. 1 and  7).The fragments for further analysis were chosen based on the position of cre in the genome.These DNA fragments were used to sequester the CcpA protein from the CcpA/DNA complex containing a labelled no-cre fragment (Fig. 7).The CcpA/ nocre complexes migrated very similarly when unla-   Unlabelled ybhE cre site (lanes 3, 5, 7, 9) and no-cre fragments (lanes 4, 6, 8, 10) were tested for their ability to sequester CcpA from labelled ybhE cre probe (10 ng) in the presence of 2 µg/ml of sonicated calf thymus DNA.

DISCUSSION
The CcpA protein plays an important role in the control of gene expression in low-GC Gram-positive bacteria (Henkin et al., 1991).CcpA is postulated to exert its regulatory functions through direct interactions with target DNA sequences, called cre sites.The majority of these data come from studies on the CcpA protein from Bacillus species, while there is no such data from lactic acid bacteria, a group of biotechnologically important bacteria.
In this work, interactions of L. lactis CcpA regulatory protein with cre sites were tested.To determine the role of CcpA in the genetic regulation in L. lactis the DNA-binding ability of this regulator to particular cre sites was studied.For the initial analysis the ybhE cre sequence, which is consistent with the cre consensus and located near the promoter region of the ybhE gene reported to be CcpAregulated (Kowalczyk et al., unpublished) was chosen.This analysis was performed by EMSA with a non-cre containing region as a negative control.For this purpose the CcpA-His-tagged protein was over-produced and purified under native conditions to obtain a biologically active protein (Kowalczyk & Bardowski, 2003).The three-dimensional structural model demonstrates that the His-tag placed at the N-terminus of the CcpA protein is located outside the protein-DNA interface, thus most probably it does not interfere with the protein/DNA interactions (Fig. 2).Additionally, CcpA-His-tagged proteins were previously used for gel retardation analysis in B. megaterium (Deutscher et al., 1995;Ramseier et al., 1995).
Binding was done in a buffer containing FBP (fructose-bisphosphate) that was previously reported to enhance specific binding of CcpA to the xyn cre sequence in B. subtilis (Galinier et al., 1999) and to be required for protection of pta cre sites of B. subtilis against DNase I digestion (Presecan- Siedel et al., 1999).The analysis of the DNA-binding ability of the CcpA itself, in the absence of the HPr-Ser-P corepressor, showed that CcpA of L. lactis was able to bind DNA in these conditions, including DNA lacking cre sites (no-cre DNA) (Fig. 3).The unspecific binding of CcpA from L. lactis was not decreased in buffer B used by Hilbrich neither was it eliminated when short oligonucleotide DNA targets were used (Fig. 4).In contrast, the CcpA protein from B. megaterium did not show unspecific binding to DNA (Aung-Hilbrich et al., 2002;Mahr et al., 2002).However, it was previously reported that the CcpA protein without its corepressor could bind to cre that flanks the transcription start site of the α-amylase structural gene, amyE in B. subtilis (Kim et al., 1995), and an unspecific interaction of CcpA with no-cre DNA was also reported in this bacterium (Miwa et al., 1994).
HPr-Ser-P was found to interact with the Histagged CcpA from B. megaterium immobilized on a Ni-NTA column (Deutscher et al., 1995) and to form a specific ternary complex with CcpA and cre (Jones et al., 1997;Schumacher et al., 2004).Circular dichroism studies showed that the DNA duplex and the CcpA dimer presented a moderate change in the DNA spectrum compared with free DNA (unspecific interaction), however, the DNA spectrum in the presence of CcpA and HPr-Ser-P exhibited a significant change consistent with a structural change in the DNA (Jones et al., 1997).Despite the strong evidence for the CcpA : HPr-Ser-P interaction, different hypotheses were proposed for the role of the CcpA corepressor in DNA binding and catabolite repression.Some studies suggest that HPr-Ser-P is necessary for specific binding of CcpA to cre (Fujita et al., 1995;Kraus et al. 1998;Aung-Hilbrich et al., 2002).However, specific binding was also observed in the absence of the corepressor in the case of PepR1, a Lactobacillus delbrueckii homologue of CcpA that specifically interacted in the absence of HPr-Ser-P with DNA binding of CcpA from Lactococcus lactis pepQ, pepX, pepI and brnQ promoter regions (Schick et al., 1999).
The presence of HPr-Ser-P of B. subtilis was found in this work to slightly enhance the CcpA binding to the ybhE cre fragment (Fig. 5).This weak effect of the protein/protein interaction could be due to the failure of cross communication between these two components of carbon catabolite repression.A similar effect was reported for the interaction of L. casei CcpA with B. megaterium HPr-Ser-P, which was 5-fold weaker than with its own HPr-Ser-P (Mahr et al., 2002).However, successful in vivo heterologous complementations were also previously shown (Davison et al., 1995;Egeter & Brückner, 1996;Monedero et al., 1997).Another possible explanation could be that HPr-Ser-P together with FBP is not sufficient to cause an increase of CcpA binding to cre sites.This could suggest that HPr-Ser-P barely increases the DNA-binding ability of the CcpA protein in L. lactis.The role of HPr-Ser-P in modulating the contact between CcpA and the cre sequence could be explained by the fact that HPr-Ser-P could stimulate the activity of the CcpA/HPr-Ser-P/cre instead of participating in the formation of the cre DNA/protein complex.
We also analysed binding of CcpA to various cre-containing DNA fragments located within the PEP:PTS regions by competition EMSA (Figs. 6 and  7).These results suggest that the ybhE, bglS, rheB, yebE, ptcB and yecA genes could be directly regulated by CcpA, which needs to be confirmed by further analysis of gene expression.
CcpA was shown not to bind specifically to three cre-containing fragments (celB cre, yebF cre, ptcA cre).The obtained results suggest that these genes are not regulated by CcpA.In the case of the ptcA gene there is no promoter upstream of ptcA and the chosen cre sequence is in fact placed in the vicinity of the gene but not of its promoter.The lack of CcpA-binding can also be explained by the fact that the ptcA gene probably forms an operon with ptcB and could be regulated by binding of CcpA to cre within the promoter region of ptcB.The lack of specific CcpA-binding to the celB cre fragment could alternatively be explained by an indirect regulation by the CcpA protein.However, as the celB gene is directly preceded by the ybhE gene and celB can be transcribed together with ybhE (T.Aleksandrzak-Piekarczyk et al., unpublished), a direct regulation by binding of CcpA to cre within the promoter region of the ybhE gene can also be suggested.
The yebE cre fragment was also found to bind CcpA with a higher affinity then the no-cre fragment.However, as the putative promoter of the yebE gene is difficult to identify, it is not possible to definitely confirm the role of this cre site in the regulation of the yebE gene.It is also likely that the yebE gene is not regulated by CcpA or can be transcribed with rheB as one operon.

Figure 1 .
Figure 1.The genetic organisation of two PEP : PTS regions in the L. lactis IL1403 genome.Sequences are derived from the L. lactis IL1403 genome (AE005176).The potential cre elements deviating from the cre consensus TGNNANCGNTNNCA by a maximum of two mismatches, but not in the central CG base pairs, were identified by sequence analysis.The cre elements that were further analysed in this study are marked as vertical black lines whereas other cre sites are marked as vertical grey lines.Genes encoding proteins homologous to the A, B and C components of the PEP:PTS system are denoted in black and those homologous to hypothetical regulators in grey.Numbering of cre sites given here refers directly to the numbering of lanes in Fig.7.Point marked as "0" represents ybhE cre.

Figure 2 .
Figure 2. Ribbon representation of the L. lactis CcpA model structure in complex with DNA and HPr (A).L. lactis CcpA DNA-binding domain (B).The N-terminus of this domain does not interact with DNA.CcpA monomers are shown in dark and light blue, DNA, green; HPr, yellow.Fragments of CcpA interacting with DNA are coloured: red, HTH motifs; orange, hinge helices.