Quantitative interdependence of coeffectors, CcpA and cre in carbon catabolite regulation of Bacillus subtilis


  • Gerald Seidel,

    1. Lehrstuhl für Mikrobiologie, Institut für Mikrobiologie, Biochemie und Genetik der Friedrich-Alexander Universität Erlangen-Nürnberg, Germany
    Search for more papers by this author
  • Marco Diel,

    1. Lehrstuhl für Mikrobiologie, Institut für Mikrobiologie, Biochemie und Genetik der Friedrich-Alexander Universität Erlangen-Nürnberg, Germany
    Search for more papers by this author
  • Norbert Fuchsbauer,

    1. Lehrstuhl für Mikrobiologie, Institut für Mikrobiologie, Biochemie und Genetik der Friedrich-Alexander Universität Erlangen-Nürnberg, Germany
    Search for more papers by this author
  • Wolfgang Hillen

    1. Lehrstuhl für Mikrobiologie, Institut für Mikrobiologie, Biochemie und Genetik der Friedrich-Alexander Universität Erlangen-Nürnberg, Germany
    Search for more papers by this author

W. Hillen, Lehrstuhl für Mikrobiologie, Institut für Mikrobiologie, Biochemie und Genetik der Friedrich-Alexander Universität Erlangen-Nürnberg, Staudtstr. 5, 91058, Erlangen, Germany
Fax: +49 9131 8528082
Tel: + 49 9131 8528081
E-mail: whillen@biologie.uni-erlangen.de


The phosphoproteins HPrSerP and CrhP are the main effectors for CcpA-mediated carbon catabolite regulation (CCR) in Bacillus subtilis. Complexes of CcpA with HPrSerP or CrhP regulate genes by binding to the catabolite responsive elements (cre). We present a quantitative analysis of HPrSerP and CrhP interaction with CcpA by surface plasmon resonance (SPR) revealing small and similar equilibrium constants of 4.8 ± 0.4 µm for HPrSerP–CcpA and 19.1 ± 2.5 µm for CrhP–CcpA complex dissociation. Forty millimolar fructose-1,6-bisphosphate (FBP) or glucose-6-phosphate (Glc6-P) increases the affinity of HPrSerP to CcpA at least twofold, but have no effect on CrhP–CcpA binding. Saturation of binding of CcpA to cre as studied by fluorescence and SPR is dependent on 50 µm of HPrSerP or > 200 µm CrhP. The rate constants of HPrSerP–CcpA–cre complex formation are ka = 3 ± 1 × 106 m−1·s−1 and kd = 2.0 ± 0.4 × 10−3·s−1, resulting in a KD of 0.6 ± 0.3 nm. FBP and Glc6-P stimulate CcpA–HPrSerP but not CcpA-CrhP binding to cre. Maximal HPrSerP-CcpA–cre complex formation in the presence of 10 mm FBP requires about 10-fold less HPrSerP. These data suggest a specific role for FBP and Glc6-P in enhancing only HPrSerP-mediated CCR.


catabolite control protein A


carbon catabolite regulation


catabolite responsive elements


catabolite repression HPr phosphorylated at serine 46










HPr kinase/phosphorylase


histidine containing protein phosphorylated at serine 46


phosphotransferase system


response units


surface plasmon resonance

Carbon catabolite regulation (CCR) in Gram-positive bacteria with low GC content is one of the most versatile regulatory processes known in bacteria. In Bacillus subtilis, the central regulator of CCR, called CcpA, represses or activates more than 300 genes involved in carbon and nitrogen utilization [1–4] and is active in the exponential but also in the stationary growth phases [5–8]. Therefore, the probably multifaceted regulatory mechanism of CcpA-mediated CCR is of considerable interest. CcpA is a member of the LacI/GalR family of bacterial regulators and binds to catabolite responsive elements (cre) in dependence of different effectors. While members of the LacI/GalR family usually respond to low molecular weight compounds, the main effectors for CcpA are the Ser46 phosphorylated histidine-containing protein (HPrSerP) and the Ser46 phosphorylated catabolite repression HPr (CrhP) [3]. HPr can also be phosphorylated at histidine 15 acting as a phosphotransferase in the phosphoenolpyruvate:sugar phosphotransferase system (PTS). In contrast, Crh residue 15 is a glutamine, which cannot be phosphorylated by the PTS. Mutation of the respective genes, ptsH and crh, results in complete loss of CCR [9–13]. HPr and Crh are phosphorylated at Ser46 by the ATP-dependent HPr kinase/phosphorylase (HPrK/P) in response to high glycolytic activity [9]. There is increasing evidence that HPrSerP and CrhP can lead to different responses. The ptsH1 mutant encoding HPr46A shows reduced CCR at many genes because CrhP substitutes only partially for HPrSerP. crh mutants, on the other hand, do not exhibit reduced CCR [14]. The properties of both effectors in CCR can depend on the growth conditions: The B. subtilis hut operon responds only to HPrSerP in Luria–Bertani medium, but to HPrSerP and CrhP in minimal medium [15]. CrhP is the sole effector for CCR of citM in minimal medium with succinate [16]. These observations may be related to the recently observed carbon source-dependent difference in crh and ptsH expression: the PTS sugars mannitol, fructose, sucrose and glucose lead to an increase of ptsH expression, whereas succinate or citrate increase crh expression [17]. CCR of some promoters, e.g. ctaBCDEF and creup dependent regulation of the gnt operon, is not affected by the ptsH1 mutation, but no data regarding the participation of CrhP are available [8,18]. No difference is observed to HPr-kinase catalysed in vitro phosphorylation of HPr and Crh [9] but the stimulatory effect of HPrSerP on CcpA binding to cre is stronger than that of CrhP [11,12]. Structural differences were revealed by NMR and X-ray indicating dimerization of Crh, but not of HPr [19–21].

Low molecular mass effectors, which would be typical inducers for members of the LacI/GalR family, are discussed controversially as effectors for CcpA. Fructose-1,6-bisphosphate (FBP) and glucose-6-phosphate (Glc6-P) enhance HPrSerP binding to CcpA [22], and FBP and NADP showed cooperative stimulation of CcpA binding to amyO in the presence of HPrSerP [23]. Glc6-P also stimulated CcpA binding to cre in the absence of HPrSerP [18,24]. Taken together, there are many observations of differential CcpA-mediated CCR, involving two phosphoproteins and several low molecular weight effectors. In an attempt to quantitatively describe the binding of HPrSerP and CrhP to CcpA, the effects of FBP and Glc6-P and their stimulation of cre binding we used surface plasmon resonance (SPR) and fluorescence to observe formation of these complexes. We describe a new role for FBP and Glc6-P in CCR because they enhance HPrSerP-mediated binding of CcpA to cre, but have no effect on CcpA–CrhP–cre interaction.


HPrSerP and CrhP binding to CcpA

SPR analyses of the protein–protein interactions of HPr, Crh and their serine phosphorylated forms with CcpA from B. subtilis have been carried out on Biacore CM5 chips, to which CcpA was covalently coupled in flowcell 2. TetR was used as control in flowcell 1 and showed no affinity for any of these proteins. Increasing concentrations (from 10 to 100 µm) of HPr or Crh did not show any binding of either protein indicating their weak affinities for CcpA. In contrast, HPrSerP or CrhP bind to CcpA under these conditions (Fig. 1). A saturation response difference of 250–280 reponse units (RU; 1000 RU, ≈ 1 ng bound ligand) was obtained for concentrations above 100 µm when a chip with 2100 RU of immobilized CcpA was used. The equilibrium constants of HPrSerP and CrhP binding to CcpA were determined by titration under steady-state conditions using 1700–2300 RU of coupled CcpA and a flow rate of 5 µL·min−1 (Supplementary material, Fig. 1). Langmuir fits of the results revealed the rather small dissociation constants of 4.8 ± 0.4 × 10−6 m for HPrSerP and 19.1 ± 2.5 × 10−6 m for CrhP. We did not detect any indication for cooperativity in the fit (Fig. 2). This result is in agreement with the hypothesis that only one form of the phosphoproteins and one interaction modus are involved in complex formation of HPrSerP or CrhP with CcpA. Structural analyses suggested that Crh may exist as a dimer at high concentrations [19–21], however, we detected only one band in 7.5% native PAGE indicating that our protein preparation contains only one form of CrhP (data not shown). Furthermore, native PAGE of phosphorylated HPrSerP or CrhP did not show any nonphosphorylated HPr or Crh (data not shown). In addition, the saturation response for CrhP is the same as that for HPrSerP bound to the same CcpA loaded chip. Since the SPR signal corresponds directly to the bound mass, and as both proteins have almost equal molecular weights, this strongly indicates binding of the same forms of HPrSerP and CrhP to CcpA. In conclusion, we assume that only the monomeric state of CrhP is present under the conditions of this study.

Figure 1.

Surface plasmon resonance analyses of effector binding to CcpA. The figure shows the sensorgrams obtained from the interaction analysis of CcpA with HPr, HPrSerP, Crh and CrhP. Dilutions (10 µm and 100 µm) of each protein were pumped at 5 µL·min−1 over a CM5 chip loaded with TetR (control) in flowcell 1 and CcpA in flowcell 2. The left diagram shows sensorgrams from injections of HPr or HPrSerP and the right diagram shows the respective sensorgrams for injections of Crh or CrhP. The concentrations of each cofactor are shown.

Figure 2.

Equilibrium titration of CcpA with HPrSerP and CrhP. The figure shows a plot of the equilibrium responses from each sensorgram vs. the corresponding HPrSerP (•) and CrhP (○) concentrations. Equilibrium constants were derived by the displayed Langmuir fits.

Effects of FBP and Glc6-P on HPrSerP- and CrhP–CcpA interaction

The effects of FBP and Glc6-P were determined by SPR at 1 µm of HPrSerP or 4 µm of CrhP so that about 20% of the immobilized CcpA is complexed. The addition of FBP or Glc6-P at millimolar concentrations led to increased complex formation of HPrSerP (Fig. 3). Titration with rising concentrations of up to 40 mm of FBP or Glc6-P did not yield saturation (Fig. 3B and D). The stimulation of HPrSerP binding to CcpA by these two effectors is highly specific because neither fructose-6-phosphate (F-6-P) nor glucose-1-phosphate (Glc1-P) showed any influence on binding (Fig. 3A and C). Titrations of CcpA with HPrSerP at 40 mm FBP or Glc6-P resulted in a KD (40 mm FBP) of 1.7 ± 0.3 × 10−6 m and a KD (40 mm Glc6-P) of 2.2 ± 0.1 × 10−6 m, respectively (data not shown). Therefore, FBP stimulates CcpA–HPrSerP complex formation at least twofold. The SPR increase at saturation is about the same for titrations with or without FBP or Glc6-P indicating that roughly the same mass binds to CcpA, ruling out a possible oligomerization of HPrSerP. Addition of only FBP or Glc6-P to the CcpA chip did not yield a signal (data not shown).

Figure 3.

Effects of low molecular mass coeffectors on HPrSerP and CrhP binding to CcpA. Sensorgrams resulting from running 1 µm of HPrSerP or 4 µm of CrhP over a CM5 chip with TetR in flowcell 1 and CcpA in flowcell 2. In addition, 5–40 mm FBP, F-6-P (A and B), Glc6-P or Glc1-P (C and D) were added. (A) The left diagram shows sensorgrams from passages of 1 µm HPrSerP with or without 40 mm FBP or F-6-P. The right diagram displays sensorgrams resulting from injections of 4 µm CrhP with or without 40 mm FBP or F-6-P. (B) The left diagram shows a titration of CcpA with mixtures of 1 µm HPrSerP and increasing concentrations (5–40 mm) of FBP. The right diagram shows the analogous titration with 4 µm CrhP instead of HPrSerP. (C) The left diagram shows sensorgrams from passages of 1 µm HPrSerP with or without 40 mm Glc6-P or Glc1-P. The right diagram shows passages of 4 µm CrhP with or without 40 mm Glc6-P or Glc1-P. (D) The left part shows sensorgrams from a titration of CcpA with mixtures of 1 µm HPrSerP and increasing concentrations (5–40 mm) of Glc6-P. The sensorgrams in the right diagram show the respective titration with 4 µm CrhP instead of HPrSerP. Analytes and their concentrations are shown in the diagrams.

CrhP binding to CcpA was not affected by FBP, Glc6-P, F-6-P or Glc1-P (Fig. 3). This result is surprising and suggests distinct functions of these phosphoproteins. Since neither nonphosphorylated HPr nor Crh interacted with CcpA in the presence of FBP, Glc6-P, F-6-P or Glc1-P in the concentration range used in these experiments (data not shown), these low molecular weight coeffectors specifically affect the HPrSerP–CcpA complex.

Stimulation of CcpA-cre complex formation by HPrSerP, CrhP, FBP and Glc6-P

The interaction of CcpA with cre was analysed by fluorescence and SPR. A C-terminally His-tagged CcpA-1W mutant carrying a single tryptophan residue at the N terminus was used for the fluorescence measurements. The regulatory activity of this mutant was determined in B. subtilis WH440 ΔccpA carrying a xynP′::lacZ fusion, transformed with either pWH1533, pWH1541 or pWH1542 expressing CcpA, His-tagged CcpA or His-tagged CcpA-1W, respectively (Table 1). The three strains expressed about the same β-galactosidase activities in dependence of the respective carbon sources (Table 2). We therefore conclude that CcpA-1W exhibits the same regulatory properties as the wild-type. CcpA-1W was prepared to homogeneity and showed increased fluorescence emission upon addition of cre DNA and HPrSerP (Fig. 4) or CrhP (data not shown). No fluorescence change was observed when HPrSerP or CrhP was added without cre DNA, or in the presence of an oligonucleotide without cre (data not shown). Thus, the fluorescence change of CcpA-1W is indicative for cre binding. No fluorescence change was observed with HPr instead of HPrSerP (Fig. 4). Titration of a CcpA-1W/cre DNA mixture with either HPrSerP (Supplementary material, Fig. 2) or CrhP (data not shown) led to increasing fluorescence, indicating complex formation of CcpA with cre. About 2.5-fold more CrhP than HPrSerP was needed to obtain the same degree of CcpA binding to cre. This result corresponds to the weaker binding of CrhP to CcpA described above. We were unable to determine the binding constant of the CcpA-1W–HPrSerP–cre complex by fluorescence, since the required low CcpA-1W concentration is below the detection limit.

Table 1.  Plasmids and strains used in this study.
Strain/plasmidCharacteristicsSource of reference
B. subtilis 168trpC2Bacillus Genetic Stock Center
B. subtilis QB7144trpC2, amyE::(xynP′-lacZ cat)[11]
B. subtilis WH440QB7144; ΔccpAThis work
B. megaterium WH419lac; gdh2Φ(xylA::lacZ); ΔccpA[28]
E. coli FT1/pLysSBL21 (DE3) ΔptsHIcrr/pLysS CmR[41]
pHT304ApR, EmR, oricolE1, ori1030[40]
pBluescript II SK +ApR, oricolE1, lacZStratagene
pWH618pBluescript II SK + ′aroA-ΔccpA-ytxDThis work
pWH1533pHT304 ccpAThis work
pWH1541pHT304 ccpAhisThis work
pWH1542pHT304 ccpA-1WhisThis work
pWH1520ApR, TcR, xylR, xylA′, oripBR, oripBc16[42]
pWH1537pWH1520 ccpAThis work
pWH1544pWH1520 ccpA-1WhisThis work
p4813ApR, ptsK[43]
pET3cApR, oripBRNovagen
pWH1576pET3c ptsH (B. megaterium)This work
pWH466pET3c ptsH (B. subtilis)This work
pWH467pET3c crh (B. subtilis)This work
Table 2.  Effect of the ccpA deletion and in trans complementation of the xynP′::lacZ fusion with wildtype ccpA and the His-tagged ccpA mutants.
Strain and (relevant genotype)β-Galactosidase activity in different media
CSKCSK + xyloseCSK + xylose + glucoseGlucose repression
QB7144 (WT)7.2 ± 1.6500 ± 103.1 ± 1.4160
WH440 pHT304 (ΔccpA)14 ± 21200 ± 501050 ± 301.1
WH440 pWH1533 (ccpA)4.8 ± 0. 4390 ± 102.0 ± 0.3200
WH440 pWH1541 (ccpAhis)5.9 ± 0.6400 ± 162.0 ± 0. 2200
WH440 pWH1542 (ccpA-1Whis)5.0 ± 1.1390 ± 113.2 ± 0.2120
Figure 4.

Fluorescence analysis of effector stimulated binding of CcpA to cre. Fluorescence emission spectra of 0.15 µm CcpA-1 W with and without HPr or HPrSerP in the presence and absence of cre are shown as indicated in the figure. Black line, 0.15 µm CcpA-1 W; dark green line, 0.15 µm CcpA-1 W with 1.5 µm HPr; red line, 0.15 µm CcpA-1 W with 1.5 µm HPrSerP; light green line, 0.15 µm CcpA-1 W with 1.5 µm HPr and 0.225 µmxylAcre; blue line, 0.15 µm CcpA-1 W with 1.5 µm HPrSerP and 0.225 µmxylAcre.

The influence of effectors on CcpA binding to cre were also analysed by SPR. About 1000 RU biotinylated 48-bp cre DNA were bound to a Biacore SA chip in flowcell 2 and 48-bp nonspecific DNA in flowcell 1 and titrated with 100 nm to 75 µm HPrSerP at 10 nm CcpA or with 100 pm to 10 nm of CcpA at 25 µm HPrSerP. The results indicated that at least 10 µm of HPrSerP and nanomolar concentrations of CcpA would have to be used for quantifications. We did not observe a steady-state response within a feasible time (data not shown), and concentrations above 5 µm of HPrSerP led to nonspecific interactions with the Biacore SA chip. Nonspecific interaction of HPrSerP or CrhP did not occur with the Biacore CM5 chip. We have used a new method to couple aminomodified DNA to that chip and measured CcpA–cre binding, HPrSerP stimulation of CcpA–cre binding and their reaction rates. Initial experiments confirmed that CcpA binds weakly to cre (data not shown) as published previously [25]. Stimulation of xylAcre binding of nanomolar concentrations of CcpA occurs only at micromolar concentrations of HPrSerP or CrhP but not with HPr or Crh. Thus, xylAcre was titrated with HPrSerP or CrhP at a fixed concentration of 10 nm of CcpA (Fig. 5A and B). The results demonstrate that 50 µm HPrSerP leads to complete saturation of cre, while the same concentration of CrhP yields only partial saturation, resembling its weaker affinity for CcpA (Fig. 2).

Figure 5.

HPrSerP and CrhP concentration dependence of the CcpA–xylAcre association rate (A) Titration of cre with HPrSerP at 10 nm of CcpA. The HPrSerP concentration for each sensorgram is shown. The baseline responses were found for 10 nm of CcpA with or without 50 µm of HPr. (B) Titration of cre with CrhP in the presence of 10 nm CcpA. The baseline response was found for 10 nm of CcpA with 50 µm of Crh. The CrhP concentration for each sensorgram is shown.

The effects of FBP and Glc6-P on cre binding were also analysed by fluorescence and SPR. Fluorescence was observed in mixtures containing 0.075 µm of CcpA-1W, 0.225 µmcre DNA and 0.3 µm HPrSerP or 0.75µMCrhP. These conditions yielded about 30% of the maximal fluorescence change, indicating partial formation of the CcpA–HPrSerP–cre complex. Titration with FBP (Fig. 6A) or Glc6-P (Supplemental Fig. 3A) yielded an increased fluorescence until saturation was reached at 2 mm FBP and 10 mm Glc6-P, repectively. This experiment showed the same fluorescence intensity obtained in the titrations with HPrSerP (Supplemental Fig. 2). We conclude that FBP or Glc6-P stimulates binding of HPrSerP to CcpA thereby increasing HPrSerP-CcpA-cre complex formation, which is monitored by flourescence. In contrast, titrations with F-6-P or Glc1-P did not result in any change of fluorescence. Replacing HPrSerP by CrhP in these titrations did not yield any stimulation of complex formation by FBP (Fig. 6B) or Glc6-P (Supplemental Fig. 3B), either, whereas the subsequent increase of the CrhP concentration resulted in complete complex formation. To verify this result by SPR, experiments using mixtures of 10 nm CcpA and 1 µm HPrSerP or 5 µm CrhP yielding partial CcpA-HPrSerP or CcpA-CrhP complex formation with cre on a CM5 chip were titrated with FBP. Fig. 6C shows the sensorgrams of both titrations. Ten millimolar FBP resulted in complete HPrSerP–CcpA–cre complex formation. In contrast, the binding of CrhP–CcpA to cre is not affected by up to 20 mm FBP. We conclude again that FBP and Glc6-P stimulate the HPrSerP-dependent binding of CcpA to cre, but have no effect on CrhP-dependent binding.

Figure 6.

Fluorescence and SPR analysis of FBP effects on HPrSerP and CrhP mediated cre binding by CcpA. (A) Plot of I/I0 (I0: CcpA-1 W fluorescence intensity only) vs. the effector concentrations for titrations of 0.075 µm CcpA-1 W and 0.225 µmcre (bsl00001), or of 0.075 µm CcpA, 0.225 µmcre and 0.3 µm HPrSerP (bsl00066) with FBP, and the titration of 0.075 µm CcpA-1 W, 0.3 µm HPrSerP and 0.225 µmcre with F-6-P (•). (B) Fluorescence titration of 0.075 µm CcpA-1 W, 0.75 µm CrhP and 0.225 µmcre with FBP. At 12 mm of FBP the CrhP concentration was raised to 3.75 µm and 5.75 µm. Points after addition of cre and corepressor are marked by arrows and labels. (C) Sensorgrams showing the influence of FBP concentrations of 2–20 mm added to 10 nm CcpA, xylAcre and 1 µm HPrSerP (left diagram) or 5 µm of CrhP instead of HPrSerP (right diagram). The baseline sensorgram results from the analysis of a mixture of 10 nm CcpA and 10 mm FBP. The FBP concentrations are shown at the right side of each sensorgram.

Association and dissociation kinetics of the CcpA–HPrSerP–xylAcre complex

Dissociation of the CcpA–HPrSerP–cre complex is very fast when buffer is injected. A much slower dissociation was observed when HPrSerP was included in that buffer (Fig. 7A). Variation of the HPrSerP concentration yielded a constant dissociation rate above 10 µm (data not shown). Since the maximal rate of CcpA–HPrSerP–cre association occurs at 50 µm of HPrSerP (see Fig. 5A) we have used this concentration to avoid bulk effects during the experiment, which are due to nonspecific signal changes caused by differences between sample composition and the running buffer. We assume that all CcpA is complexed with HPrSerP under these conditions. Therefore, the association and dissociation rate constants of the HPrSerP–CcpA–cre complex were determined with 50 µm HPrSerP in all buffers and increasing concentrations from 1 to 30 nm of CcpA. We fitted the sensorgrams according to the 1 : 1 Langmuir binding model implemented in the biaevaluation 3.1 software, assuming association and dissociation of the CcpA–HPrSerP complex from cre under these conditions. The respective sensorgrams and fits for the rate constants are shown in Fig. 7B, yielding a ka of 3 ± 1 × 106 m−1·s−1 and a kd of 2.0 ± 0.4 × 10−3 s−1 resulting in an apparent KD of 6 ± 3 × 10−10 m at an average deviation χ2ass. = 3–4 and χ2diss. = 1. The sensorgrams from a titration of cre with increasing concentrations of CcpA in the presence of 5 µm HPrSerP and 10 mm FBP are shown in Fig. 7C. The same fitting as above assuming association or dissociation of a CcpA–HPrSerP–FBP complex from cre yields the constants ka = 2.2 ± 0.5 × 106 m−1·s−1 and kd = 2.7 ± 0.8 × 10−3 s−1 resulting in an apparent KD of 1.2 ± 0.4 × 10−9 m at χass2 = 2–3 and inline image = 1. These constants are very similar to the ones obtained without FBP at 50 µm HPrSerP suggesting that FBP decreases the amount of HPrSerP necessary for complete binding of CcpA to cre.

Figure 7.

Kinetic analysis of CcpA-HPrSerP binding to xylAcre by SPR. (A) Sensorgrams of mixtures containing 10 nm CcpA and 5 µm HPrSerP are shown. In the red sensorgram the dissociation is observed in running buffer and in the blue sensorgram the dissociation is observed first in running buffer with and then without 5 µm HPrSerP as indicated. (B) The rate constants were obtained from titrations of xylAcre with mixtures of 1–30 nmB. subtilis CcpA and 50 µm HPrSerP (running buffer with 50 µm HPrSerP) or (C) 1–30 nmB. subtilis CcpA, 5 µm HPrSerP and 10 mm FBP (running buffer with 5 µm HPrSerP and 10 mm FBP). The concentrations of the CcpA–HPrSerP are assumed to be the same as those of CcpA and are depicted in the respective colour. The fits of the association phases are drawn as black lines.


Many qualitative and some quantitative studies of various effector molecules affecting CcpA–cre interaction have led to a general mechanism of action for CCR in B. subtilis[11,12,23,25,26]. However, the current model does not explain all results, e.g. it is not clear how similar PTS sugars such as glucose, fructose or mannitol lead to quite different extents of CCR, and how carbon sources like glucitol or succinate lead to ptsH- or crh-dependent CCR [9,11,16,27,28]. The different roles of HPr and Crh in CcpA-mediated CCR are particularly mysterious, since their phosphorylation by HPrK/P is similarly effective [9,29]. CrhP may be specifically active in CCR brought about by nonsugar compounds as described for citM[16]. The approximately fourfold stronger affinity of HPrSerP for CcpA compared to CrhP found here may contribute to the weaker stimulation of CrhP for CcpA binding to xylAcre, glpFKcre, ptacre and xynPcre[11,12,30], but it seems likely that other factors also contribute to differential regulation. For example, the KD of the B. subtilis HPrSerP–CcpA complex of ≈ 5 µm is almost identical to that determined for the respective Lactobacillus casei proteins (4 µm), but despite the fact that B. subtilis- and B. megaterium-derived HPrSerP showed the same fivefold lower KD for binding to L. casei CcpA, only the B. subtilis but not the B. megaterium ccpA mutant can be complemented by L. casei ccpA[26]. This indicates that CcpA–cre complex formation may be influenced by more factors than the CcpA–HPrSerP affinity.

Stimulation of HPrSerP–CcpA complex formation by FBP and Glc6-P has been observed qualitatively before [22]. Footprinting has indicated that HPrSerP and CrhP mediated CcpA–cre complex formation are stimulated by FBP [11,12]. The data presented here establish for the first time distinct mechanisms for these two effectors as only HPrSerP binding to CcpA responds to the coeffectors FBP and Glc6-P. Consequently, only the HPrSerP–CcpA–xylAcre interaction is stimulated by FBP and Glc6-P, but not CrhP–CcpA–xylAcre complex formation. The stimulatory concentrations of approximately 10 mm FBP or Glc6-P are within the range of physiological variance of these compounds [31,32]. FBP or Glc6-P reduce the concentration of HPrSerP necessary for complete occupation of CcpA, and, in turn, 10 mm FBP leads to an approximately tenfold reduction of the amount of HPrSerP necessary for complete occupation of cre by CcpA–HPrSerP. Thus, in the presence of these mediators at least 40-fold more CrhP compared to HPrSerP would be necessary to mediate full repression. These properties could explain the ptsH-specific CCR in the presence of glucitol [11] because this non-PTS sugar is converted to FBP [33]. Furthermore, the stimulatory effect of Glc6-P could explain the stronger CCR exerted by glucose as compared to other PTS sugars [9,27,28]. Crh-mediated CCR occurs in the presence of succinate and glutamate [16]. Since this is a physiological situation with low intracellular concentrations of Glc6-P and FBP, there may be yet unknown effectors for CcpA.

The equilibrium constants of HPrSerP and CrhP binding to CcpA from B. subtilis are quite low, but they are very well adjusted to the cellular concentrations of 1 µm of CcpA and 0.1–2 mm of HPrSerP, as found in Bacilli and Streptococci in the presence of glucose [34,35]. The low affinity of HPrSerP to CcpA makes the in vitro analysis of the coupled binding to cre difficult. This explains the unusually high concentrations of CcpA that had to be used to detect DNA binding in all previous studies, except for binding to amyO[23] and rocGcre[13]. We have previously determined a low apparent equilibrium constant of KD = 200 nm for the CcpA–HPrSerP–cre complex from B. megaterium by EMSA and SPR [25], because we assumed a KD of at least 500 nm for the CcpA–HPrSerP complex and consequently used not enough HPrSerP to obtain saturation of CcpA. These conditions also masked the effects of FBP and Glc6-P. The KD of the CcpA–HPrSerP complex and the titrations of cre with HPrSerP at a constant CcpA concentration determined here show that at least 50 µm of HPrSerP is required to assure complete complex formation of HPrSerP with CcpA, a prerequisite for quantification of the CcpA–HPrSerP–cre interaction.

The rate and equilibrium constants determined here agree well with those determined for other members of the LacI/GalR family of bacterial regulators, like PurR in the presence of guanine (ka = 1.5 ± 2 × 107 m−1·s−1; kd = 1.2 ± 0.2 × 10−3 s−1; KD = 0.8 ± 1 × 10−10 m) [36] and LacI (ka = 2 × 106 m−1·s−1; kd = 3.5 × 10−4 s−1; KD = 2 × 10−10 m) [37]. However, there may be two different types of CcpA–cre interactions. CcpA binding to the cre sites at the xylA, xynP[11], pta[12], glpFK[30] or gnt[38] promoters is very weak or not detectable without cofactors, whereas binding to amyO[23] or rocGcre[13] is strong. HPrSerP at 0.68 µm stimulated CcpA binding to amyO only 10-fold and 2 mm FBP with 0.68 µm HPrSerP stimulated it 300-fold, whereas CcpA–xylAcre binding is stimulated at least 1000-fold in the presence of 50 µm HPrSerP [23]. Thus, different cre sequences found in many genes or operons may respond in a differential manner to FBP- or Glc6-P-mediated stimulation.

Experimental procedures

Plasmid construction and bacterial strains

Strains and plasmids used in this study are listed in Table 1. For in frame deletion of ccpA in B. subtilis two DNA fragments were amplified from chromosomal DNA from B. subtilis 168, where primer pairs dccpA1 (5′-ATAATAATAGAGCTCGCTGTGCCGATTTTGAAACAAG-3′) and dccpA2 (5′-TATTATTATAGCGGCCGCAATATTGCTCATCCTAAAACC-3′) yielded fragment 1 and dccpA3 (5′-ATAATAATAGCGGCCGCTGAAGCACTGCAGCATCTGATG-3′) with dccpA4 (5′-TATTATTATGGTACCTTTTCGGTGCCGTTCCTCC-3′) yielded fragment 2. Fragment 1 comprises the sequence from 500 bp upstream to 13 bp downstream of ccpA translational start with SacI and NotI restriction sites at the 3′- and 5′-termini, respectively. Fragment 2 includes the ccpA sequence from basepair 684–438 bp of ytxD downstream from ccpA with NotI and KpnI restriction sites at the 3′- or 5′- ends, respectively. Plasmid pWH618 was constructed by cloning these fragments into pBluescriptII SK + via the restriction sites SacI and NotI for fragment 1 and NotI and KpnI for fragment 2. The product carries a ccpA fragment lacking bases 13–684 (corresponding to residues Thr5–Leu228). The NotI restriction site was positioned between bases 13 and 684 resulting in a linker with three alanines replacing CcpA residues 6–227. Strain WH440 was generated by cotransformation of B. subtilis 168 with the plasmid pWH618 and chromosomal DNA from B. subtilis QB7144 (xynP′::lacZ). Transformants were selected for the presence of the cat resistance gene linked to the xynP′::lacZ fusion from QB7144 on CSK medium supplemented with 1% glucose, 0.2% xylose and 80 mg·mL−1 X-Gal and 5 mg·mL−1 chloramphenicol. Blue stained colonies showing deregulation of xynP′::lacZ were picked for verification of the deletion of ccpA by Western blotting. For complementation of WH440 ccpA was amplified with the primer pair ccpAmut1 (5′-ATAATATCTAGAACCAAGTATACGTTTTCATC-3′) and ccpAstd2 (5′-TATTATTATGGATCCTTTTCTTATGACTTGGTTT-3′). This fragment contains the ccpA promoter 290 bp upstream from the start codon [39]. This fragment was cloned into the shuttle vector pHT304 [40] via the restriction sites XbaI and BamHI resulting in pWH1533. For construction of the vector pWH1541 ccpA was amplified by ccpAmut1 and ccpAnot (5′-TATTATTATGCGGCCGCTGACTTGGTTGACTTTCTA-3′) using pWH1533 as template. The His-tag encoding sequence was amplified by primers hisnot (5′-ATAATAGCGGCCGCGGGCGGTCATCACCATCACCATCACTA-3′) and hisbam (5′-TATTATTATGGATCCTTAGCTTCCTTAGCTCCTGA-3′) from vector pQE17. After restriction of the ccpA fragment with XbaI and NotI and the His-tag encoding fragment with NotI and BamHI, both were cloned in a three-armed ligation into pHT304 via the restriction sites and XbaI and BamHI. For construction of pWH1542 ccpA was mutagenized via two-step mutagenesis using primers ccpAmut1, hisbam and ccpA +1W (5′-CGTAATATTGCTCCACATCCTAAAACC-3′). The resulting fragment encoding C-terminally His-tagged ccpA carrying an additional tryptophan residue at the N terminus was cloned into pHT304 via XbaI and BamHI. For overexpression of HPr and Crh from B. subtilis or HPr from B. megaterium either ptsH genes or crh were cloned into pET3c via NdeI and BamHI resulting in pWH466, pWH467 and pWH1576. For overexpression Escherichia coli FT1 [41] was transformed with the latter plasmids. For overexpression ccpA from B. subtilis was subcloned from pWH1533 into pWH1520 resulting in pWH1537. By analogy ccpA-1Whis was subcloned from pWH1542 yielding pWH1544. B. megaterium WH419 overexpressed either proteins after transformation with pWH1537 or pWH1544.

β-Galactosidase assays

Cells for β-galactosidase assays were grown overnight at 37 °C in CSK minimal medium. From overnight cultures the same medium and CSK supplemented with 0.2% xylose or with 1% glucose, 0.2% xylose were inocculated to D600 = 0.1 and grown at 37 °C until a D600 value of ≈ 0.4 was reached. One-hundred microlitres bacterial culture were diluted with 900 µL Z-buffer (60 mm Na2HPO4, 40 mm NaH2PO4, 10 mm KCl, 1 mm MgSO4, 50 mmβ-mercaptoethanol, pH 7). After lysis with lysozyme and Triton-X-100 β-galactosidase activity was determined as described earlier [28].

Preparation of proteins

CcpA from B. subtilis was expressed in B. megaterium WH419/pWH1537 and C-terminally His-tagged CcpA-1 W in B. megaterium WH419/pWH1544 (Table 1) as described [24]. For purification the cells were disrupted by ultrasonification, centrifuged for 45 min at 48 400 g at 4 °C and incubated with 5 µg·mL−1 RNaseA and 10 µg·mL−1 DNaseI (Sigma, Munich, Germany). Wild-type CcpA was purified by subsequent cation exchange chromatography on POROS 20 HS (Perseptive Biosystems, Framingham, MA, USA), desalting (Pharmacia Biotech, Freiburg, Baden Wuerttemberg, Germany), anionic exchange chromatography on Fractogel EMD TMAE (Merck, Darmstadt, Hesse, Germany) and gelfiltration on Superdex G75 (Pharmacia Biotech). C-terminally His-tagged CcpA-1 W was purified using Ni-affinity chromatography on POROS 20 MC (Perseptive Biosystems). Further purification was achieved by gelfiltration on Superdex G75 (Pharmacia Biotech). HPr from B. megaterium was overproduced in E. coli FT1/pWH1576, HPr from B. subtilis in E. coli FT1/pWH466 and Crh in E. coli FT1/pWH467. The crude lysates have been incubated with 5 µg·mL−1 RNaseA and 10 µg·mL−1 DNaseI (Sigma), then prepurified by heat denaturation for 20 min at 70 °C and 65 °C, respectively. After centrifugation from the precipitated proteins, HPr or Crh could be extracted from the supernatant. Phosphorylation of either protein was performed in the prepurified crude lysate using a HPr kinase extract as described [25]. Purification of HPr, HPrSerP, Crh or CrhP was achieved by anion exchange chromatography on DEAE Sephacel (Pharmacia Biotech) and subsequent gelfiltration on Superdex G75 (Pharmacia Biotech). The activities of both phosphorylated proteins was assumed to be 100%, as there is no obvious method for activity determination. However, we assume that the potentially active fractions are the same for both, similarly heat stable proteins. Therefore, the ratio of constants should be reliable.

Determination of protein concentration

Protein concentrations were measured using a Bio-Rad (Munich, Germany) Bradford dye binding assay. BSA was used as a standard. The concentration of purified CcpA-1 W protein was confirmed by UV spectroscopy at 280 nm using an extinction coefficient of ε280nm = 21300 m−1·cm−1.

Preparation of cre DNA

Forty-eight nucleotide synthetic oligonucleotides containing xylAcre (forward: 5′-CTAATAAAATTAATCATTTTGAAAGCGCAAACAAAGTTTTATACGAAG-3′; backward: 5′-CTTCGTATAAAACTTTGTTTGCGCTTTCAAAATGATTAATTTTATTAG-3′) and 26-nt oligonucleotides containing xylAcre (forward: 5′-AATCATTTTGAAAGCGCAAACAAAGT-3′; backward: 5′-ACTTTGTTTGCGCTTTCAAAATGATT-3′) or a nonspecific DNA sequence (5′-AATCATTTATGGCATAGGCAACAAGT-3′; backward: 5′-ACTTGTTGCCTATGCCATAAATGATT-3′) were hybridized and used for analyses without further purification. Both forward 26-nt oligonucleotides carry a C6 aminolinker at the 5′-end. All oligonucleotides were purchased with or without modification at MWG Biotech (Ebersberg, Germany). The concentration of the hybridized DNA was determined using an extinction coefficient of ε = 1186 × 10−6 m−1·cm−1 as determined from the nucleotide composition.

SPR measurements

SPR measurements with CcpA, HPrSerP or CrhP each from B. subtilis or xylAcre, were analysed using a BIAcoreX instrument operated at 25 °C (BIAcoreX, Uppsala, Sweden). For the analysis of protein–protein interactions CcpA was immobilized by amine coupling on the carboxylated dextran matrix of a CM5 sensorchip (Biacore AB) in flowcell Fowcell 1 contained TetR from E. coli and was used as a reference. For immobilization on the activated chip matrix (injection of 35 µL of a mixture containing 50 mmN-hydroxysuccinimide and 200 mmN-ethyl-N′-(3-dimethylaminopropyl)carbodiimide hydrochloride in desalted, sterile water) the proteins were injected at 500 nm concentrations in 10 mm sodiumacetate, pH 5. After coupling of the proteins the residual activated carboxyl groups were deactivated by injection of 1 m ethanolamine hydrochloride/NaOH, pH 8.5. Both proteins, CcpA and TetR-B/D, were adjusted to equal immobilization levels of 1700–2100 RU on different sensorchips. During immobilization and interaction analyses HBS/EP buffer (0.01 m Hepes pH 7.4, 0.15 m NaCl, 3 mm EDTA, 0.005% polysorbate) purchased from Biacore was used as a running buffer at a flowrate of 5 µL·min−1. For the interaction analyses, the injected analyte volume was adjusted to the amount needed for a constant response difference indicating the equilibrium of interaction of CcpA with HPrSerP or CrhP. The concentration of the complex is measured directly as the steady state response [R(eq)] in SPR. As the analyte is constantly replenished during sample injection, the concentration of free analyte is equal to the bulk analyte concentration. The equilibrium constants were determined by Langmuir fits of plots from the steady state response vs. the analyte concentrations. Evaluation was done using the Langmuir equation for 1 : 1 ligand binding of the program sigmaplot™8.0 (SPSS Inc., Chicago, IL, USA). Each equilibrium constant and deviations were determined from three different titrations. For interaction analyses of CcpA with xylAcre we immobilized amino-modified 26-meric DNA (see preparation of cre DNA) containing the xylAcre or a nonspecific DNA sequence on the surface of Biacore CM5 chips. We used a new method for coupling of amino-modified DNA to Biacore CM5 chips. This method uses cetyltrimethylammoniumbromide (CTAB) micelles as carriers to immobilize DNA on the carboxymethylated dextran matrix (H. Sjöbom, Biacore AB, Uppsala, Sweden, personal communication). We coupled hybridized nonspecific DNA in flowcell 1 and xylAcre containing DNA in flowcell 2 by injection of mixtures containing 5 µm of amino-modified DNA, 0.6 mm CTAB in 10 mm Hepes at a pH of 7.4 over a CM5 chip that was activated as described above. During coupling we used HBS-N (10 mm Hepes, 150 mm NaCl) as a running buffer at a flow rate of 5 µL·min−1. After deactivation of residual activated carboxyl groups as described above ≈ 280 RU DNA remained stably attached to the chip, but only ≈ 30–60 RU were functional as calculated from the maximum response of CcpA-HPrSerP binding to xylAcre. For all CcpA–cre interaction analyses HBS-EP buffer purchased from Biacore AB was used as a running buffer. The mass transport limitation was tested by alteration of flow rates. A flow rate of 40 µL·min−1 was suitable for all experiments to minimize mass transport. To regenerate the chip surface the dissociation of the CcpA–HPrSer46P complex was stopped by injection of 80 µL HBS-EP buffer at 40 µL·min−1 after each injection. Fits showed that concentrations > 30 nm CcpA or CcpA–HPrSerP complex, which saturate the cre coupled to the chip, result in biphasic sensorgrams. We analysed only sensorgrams from 1 nm to 30 nm CcpA in the presence of HPrSerP or HPrSerP and FBP. The titrations for the kinetic measurements have been carried out twice for each protein complex, CcpA–HPrSerP or CcpA–HPrSerP–FBP. FBP (Fluka) F-6-P, Glc6-P or Glc1-P (Sigma) were diluted immediately before each experiment in HBS-EP buffer to 100 mm stock solutions and if necessary adjusted to pH 7.4. In order to prevent bulk effects the HBS-EP running buffer was adjusted to the concentration of these compounds and then supplied with HPrSerP if required.

Fluorescence measurements

For fluorescence spectroscopic measurements we used CcpA-1 W and Crh or CrhP from B. subtilis and xylAcre and HPr or HPrSerP from B. megaterium. FBP, F-6-P, Glc6-P and Glc1-P were from Fluka or Sigma. All components were dissolved in 10 mm Tris/HCl pH 7.5, 150 mm NaCl, 5 mm MgCl2 and 1 mm dithiothreitol. Fluorescence emission spectra were recorded on a Spex Fluorolog 3 spectrofluorimeter at a slit width set of 4 mm. Tryptophan emission was excited at a wavelength of 295 nm and a temperature of 25 °C. In titration experiments, fluorescence emission was scanned in a wavelength range of 340–370 nm. As a measure for fluorescence intensity, peak areas were determined by integration with computer software Grams/32 from Galactic Industries Corporation. For evaluation the intensities were corrected for increasing sample volume and inner filter effect at each titration step by the mathematical transformation:


where Iexp and I are the fluorescence intensity before and after correction. V0 and V are the sample volumes at the start of the experiment and at each titration step. A295 is the overall absorption of the sample at the excitation wavelength, which was measured for each titration step at a Pharmacia Biotech Ultrospec 4000.


We thank Dr Lwin Mar Aung-Hilbrich for providing pWH1576 and Eva Maria Henßler for TetR. We are grateful to Dr Hans Sjöbom and Dr Anders Sjödin from Biacore AB for their help with CTAB mediated DNA coupling. This work was supported by the Deutsche Forschungsgemeinschaft through SFB 473 ‘Schaltvorgänge der Transkription’, the Graduiertenkolleg 805 and the Fonds der Chemischen Industrie.

Supplementary material

The following material is available from http://www.blackwellpublishing.com/products/journals/suppmat/EJB/EJB4682/EJB4682sm.htm