Catabolite control of Escherichia coli regulatory protein BglG activity by antagonistically acting phosphorylations
1999; Springer Nature; Volume: 18; Issue: 12 Linguagem: Inglês
10.1093/emboj/18.12.3370
ISSN1460-2075
Autores Tópico(s)Protein Structure and Dynamics
ResumoArticle15 June 1999free access Catabolite control of Escherichia coli regulatory protein BglG activity by antagonistically acting phosphorylations Boris Görke Boris Görke Institut für Biologie III, Universität, Schänzlestrasse 1, D-79104 Freiburg, Germany Search for more papers by this author Bodo Rak Corresponding Author Bodo Rak Institut für Biologie III, Universität, Schänzlestrasse 1, D-79104 Freiburg, Germany Search for more papers by this author Boris Görke Boris Görke Institut für Biologie III, Universität, Schänzlestrasse 1, D-79104 Freiburg, Germany Search for more papers by this author Bodo Rak Corresponding Author Bodo Rak Institut für Biologie III, Universität, Schänzlestrasse 1, D-79104 Freiburg, Germany Search for more papers by this author Author Information Boris Görke1 and Bodo Rak 1 1Institut für Biologie III, Universität, Schänzlestrasse 1, D-79104 Freiburg, Germany *Corresponding author. E-mail: [email protected] The EMBO Journal (1999)18:3370-3379https://doi.org/10.1093/emboj/18.12.3370 PDFDownload PDF of article text and main figures. ToolsAdd to favoritesDownload CitationsTrack CitationsPermissions ShareFacebookTwitterLinked InMendeleyWechatReddit Figures & Info In bacteria various sugars are taken up and concomitantly phosphorylated by sugar-specific enzymes II (EII) of the phosphoenolpyruvate:sugar phosphotransferase system (PTS). The phosphoryl groups are donated by the phosphocarrier protein HPr. BglG, the positively acting regulatory protein of the Escherichia coli bgl (β-glucoside utilization) operon, is known to be negatively regulated by reversible phosphorylation catalyzed by the membrane spanning β-glucoside-specific EIIBgl. Here we present evidence that in addition BglG must be phosphorylated by HPr at a distinct site to gain activity. Our data suggest that this second, shortcut route of phosphorylation is used to monitor the state of the various PTS sugar availabilities in order to hierarchically tune expression of the bgl operon in a physiologically meaningful way. Thus, the PTS may represent a highly integrated signal transduction network in carbon catabolite control. Introduction Although it was initially thought that different mechanisms of signal transfer may exist for the three kingdoms of life, it has recently been discovered that eukaryotic-type protein kinases are also found in prokaryotes (for a review see Zhang, 1996), whereas two-component systems, which were thought to be restricted to bacteria, were found to be involved in eukaryotic signaling (for reviews see Loomis et al., 1997; Wurgler-Murphy and Saito, 1997). In contrast, the phosphoenolpyruvate:sugar phosphotransferase system (PTS), which is ubiquitous in bacteria, seems to be bacteria specific. The PTS is not only central in uptake and phosphorylation of various sugars, but also mediates chemotaxis towards these sugars, as well as transcriptional regulation of numerous genes (for reviews see Postma et al., 1993; Saier and Reizer, 1994). The PTS catalyzes phosphotransfer reactions from phosphoenolpyruvate to various sugars, a process which involves at least three proteins. First, enzyme I (EI) autophosphorylates with phosphoenolpyruvate, and then donates the phosphoryl group to histidine protein (HPr), which in turn phosphorylates the respective sugar-specific transport proteins enzymes II (EII) from which phosphate is transferred to the incoming sugar concomitantly with its transport. EIIs consist of at least three functionally different domains, IIA, IIB and IIC, which may be present on the same polypeptide chain or may be encoded by different genes. The phosphocarrier protein HPr phosphorylates the respective EIIs at a histidine residue in the IIA domain, from which phosphate is transferred to a residue of the IIB domain, in most cases a cysteine, and then to the sugar. Membrane bound domain, IIC, forms the translocation channel and the sugar-specific substrate-binding site. The PTS is also central in carbon catabolite repression, exerted in Gram-negative bacteria by the glucose-specific IIAGlc (formerly EIIIGlc). Phosphorylated IIAGlc is thought to activate adenylate cyclase, whereas non-phosphorylated IIAGlc binds and inhibits various proteins concerned with uptake and metabolism of non-PTS sugars. This mechanism may, therefore, explain preferred utilization of some PTS carbohydrates over non-PTS carbohydrates. Preferential utilization of one PTS sugar over another has also been observed several times, but the cause of this hierarchical utilization remained unclear (Postma et al., 1993). The transcriptional regulator BglG from Escherichia coli regulates expression of the bgl operon, encoding proteins concerned with regulated uptake and utilization of β-glucosidic sugars (Prasad and Schaefler, 1974; Schnetz et al., 1987). BglG was the first identified member of a fast-growing family of related proteins shown to be directly regulated by the PTS in a substrate-dependent manner (Rutberg, 1997; Tortosa et al., 1997; Stülke et al., 1998). BglG is the product of the first gene of the operon and acts as a transcriptional antiterminator at two ρ-independent terminators within the operon (Mahadevan and Wright, 1987; Schnetz and Rak, 1988). The second gene of the operon, bglF, encodes the β-glucoside-specific EIIBgl, which is a member of the PTS-family of sugar transport proteins. EIIBgl has also been shown to be a negative regulator of BglG (Mahadevan et al., 1987). In the absence of substrate, when phosphate cannot be transferred to the sugar, EIIBgl phosphorylates and concomitantly inactivates BglG, thereby preventing transcription of the operon. When β-glucosides become available, EIIBgl dephosphorylates and thereby reactivates BglG (Amster-Choder et al., 1989; Amster-Choder and Wright, 1990; Schnetz and Rak, 1990). In this report we re-investigated regulation of BglG activity. Data presented show that to be active, BglG requires the presence of the general PTS proteins, EI and HPr. Complementation analysis revealed that phospho-HPr is necessary for activation and that the diphosphoryl transfer protein (DTP), which shares an HPr-like domain, can substitute for HPr in this process. Furthermore, we show that BglG is phosphorylated in vivo in a PTS-dependent reaction also in the absence of EIIBgl, suggesting that HPr activates BglG by phosphorylation and that two different and antagonistically acting phosphorylation reactions control BglG activity. When we investigated the possible role of this positively acting HPr-mediated phosphorylation, we found that this mechanism may function to downregulate BglG activity and thus bgl operon expression when other, more favorable PTS sugars become available, by draining the activating phosphoryl groups away from BglG. This control, which makes use of the reversible flow and limited pool of transferable phosphoryl groups within the PTS, represents an unprecedented level of carbon catabolite control in E.coli. Results BglG requires the PTS for its activity To determine antitermination activity of BglG, we made use of our antitermination reporter plasmid described previously (Schnetz et al., 1996). In this plasmid (Figure 1A) the lacZ reporter gene is preceded by terminator t2 of the bgl operon (bgl-t2) which in the absence of the active BglG protein blocks elongation of transcription initiated at constitutive promoter P16. Also present on the plasmid is the lacIq gene providing the lac repressor for the controlled expression of bglG which is under tacOP control on a second plasmid (Figure 1B). In the absence of the bglG-containing plasmid, 37 units of β-galactosidase activity were synthesized in a wild type strain background as well as in an isogenic background carrying a deletion (Δpts) which encompasses genes ptsH, ptsI and crr encoding HPr, EI and IIAGlc, respectively. This low activity reflects the leakiness of terminator t2 (Schnetz and Rak, 1988; dotted line in Figure 1B). In the presence of the bglG expression plasmid, and with increasing concentrations of isopropyl-β-D-thiogalactopyranoside (IPTG) as inducer for BglG synthesis, enzyme activity increased in the wild type background from 177 units (no IPTG) to 1776 units (2 mM IPTG). In the pts deletion strain, enzyme activity did not start to rise above background at concentrations below 0.05 mM IPTG (49 units) and only reached 676 units at saturating concentrations (2 mM). Moreover, the ratio of enzyme activities (wt/Δpts) decreased with increasing IPTG concentrations from 14.1 (0.05 mM) to 2.6 (2 mM), indicating that BglG requires an intact PTS for full activity and that this requirement can be partially overcome by overproduction of BglG. To support this conclusion we verified that deletion of the pts operon does not negatively influence gene expression within our detection system. To this end we employed transformants carrying a plasmid (pFDY225), which is identical to the antiterminator test plasmid (Figure 1A) but lacks the bgl terminator. In addition we employed a plasmid (pFDX3549), which is isogenic to the BglG expression plasmid (Figure 1B) but carries the lacZ gene instead of bglG. Indeed, enzyme activities directed by these plasmids in the wild type strain were not significantly different from those synthesized in the isogenic pts deletion derivative (data not shown). Figure 1.BglG requires an intact PTS to be active, but its overproduction can partially overcome this requirement. Antitermination reporter plasmid pFDY226 (A) was cotransformed with bglG expression plasmid pFDX2469 (B) into E.coli strain R1279 (pts+; white histograms) and its isogenic Δpts derivative R1653 (shaded histograms). Increasing levels of BglG synthesis were driven by increasing concentrations of inducer (IPTG) and β-galactosidase activity was determined as a measure for antitermination activity. The ratio of enzyme activity synthesized by the two strains is given as filled triangles. For IPTG concentrations below 50 μM enzyme activity remained at background levels with the Δpts strain which was 37 units for both strains when transformed with the antitermination test plasmid alone (dotted line). Download figure Download PowerPoint Activation of BglG requires phospho-HPr or phospho-DTP Next we wanted to define in more detail the requirements for activation of BglG by the PTS. To avoid the problems imposed by overproduction of BglG due to high gene dosage with multi-copy plasmids, we inserted the tacOP-driven bglG expression cassette either with (Figure 2B) or without the bglF gene (encoding EIIBgl) (Figure 2C) into the chromosome. The different strain backgrounds used together with complementing plasmids are given in Figure 2A. As can be seen from the data, β-galactosidase activity was low in all cases when IPTG was omitted. As expected, it remained low in the presence of IPTG when bglF was coexpressed together with bglG in the wild type strain and increased to 325 units when salicin was added as substrate for bglF-encoded EIIBgl (Figure 2B, column 1) reflecting the relief from negative control exerted by the phosphorylated EIIBgl on BglG activity (see Introduction). Likewise, BglG activity was constitutive and high in the absence of bglF (Figure 2C, column 1). In the Δpts background (columns 2), β-galactosidase activities were reduced to background levels verifying the PTS dependence of BglG activity demonstrated above. Activation of BglG as well as its negative regulation were fully complemented when HPr together with EI were expressed from a plasmid (columns 3). Expression of HPr alone did not result in complementation (column 4), whereas expression of EI alone led to partial complementation (columns 5). HPr thus appears not to be absolutely required for activation and negative control of BglG. However, it has been reported that a defect in HPr can be complemented for PTS sugar transport in mutants defective for fruR (Saier and Ramseier, 1996). The repressor FruR negatively controls the fructose inducible fru operon which codes for fructose-specific PTS functions and a fructose-1-phosphate kinase. Responsible for the complementation is the product of the fruB gene, DTP (diphosphoryl transfer protein), which carries a domain homologous to HPr. DTP thus was a good candidate which possibly could substitute for HPr in BglG control. We therefore repeated the experiment shown in columns 5 with a culture pregrown in fructose as inducer of the fru operon. This pretreatment indeed resulted in better complementation of activation as well as negative control of BglG activity (columns 6). Consequently, we analyzed the involvement of fru operon functions. Deletion of the operon in the wild type strain had no appreciable effect (compare columns 1 and 7). We next crossed in the pts deletion resulting in the double-deletion strain Δpts,Δfru. As expected, enzyme activities were equally low in this strain as in the strain carrying the pts deletion (compare columns 2 and 8). They remained low when the defects were complemented by a ptsH (HPr) or ptsI (EI) expression plasmid (see columns 9 and 10). We can thus conclude that restoration of BglG activity and its negative control in the pts deletion strain expressing EI (columns 5 and 6) is due to a function encoded by the fru operon. Since the likely candidate is DTP, we complemented the double-deletion strain with an artificial operon encompassing fruB (DTP) and ptsI(EI). As can be seen from the data in column 11, the activity of BglG as well as its negative control were fully restored. It was likewise restored when the strain expressed ptsH and ptsI (columns 12), but not when the ptsH gene was replaced by an allele carrying a point mutation within the phosphorylation site (columns 13). Together, the data indicate that BglG is activated by the PTS and that this activation requires phosphorylated HPr, which can be substituted by the phosphorylated form of DTP. Figure 2.BglG requires phospho-HPr for activation which can be substituted by phospho-DTP. Genes bglG and bglF (B) or bglG alone (C) under tacOP control were integrated into the chromosome by site-specific recombination to obtain single-copy situations and subsequently crossed into various strain backgrounds harboring different expression plasmids (A) also driven by tacOP and thus inducible by IPTG. In addition, antitermination reporter plasmids pFDX2676 or pFDX3158 were present. These plasmids are identical to plasmid pFDY226 (Figure 1A) but carry, for compatibility reasons, different antibiotic resistance markers. Inducers were added as indicated in the figure and β-galactosidase activities were determined as described in Materials and methods with the exception of lane 6. In these experiments, cells were pregrown in M9-fructose medium which was substituted by M9-glycerol at the time point when inducers were added. This was necessary to avoid interference of BglG activity due to transport of fructose (a PTS-sugar) (see data below and Discussion). The following strains were employed: in (B) lane 1, R1752; lanes 2–6, R2013; lane 7, R1971; lanes 8–13, R1977. In (C) lane 1, R1958; lanes 2–6, R2051; lane 7, R1974; lanes 8–13, R1979. These strains harbored the following plasmids: (B and C) lanes 1 and 2 and 7 and 8, pFDX2676; lane 3, pFDX2676/pFDX3155; lanes 4 and 9, pFDX3158/pFDX3160; lanes 5, 6 and 10, pFDX3158/pFDX3161; lane 11, pFDX3158/pFDX3221; lane 12, pFDX2676/pFDX3155; lane 13, pFDX3158/pFDX3223. Download figure Download PowerPoint BglG is phosphorylated by the PTS even in the absence of EIIBgl The requirement for phosphorylated HPr suggested that activation of BglG might involve HPr-catalyzed phosphorylation of BglG. This possibility was studied by in vivo phosphorylation of BglG. Strains carrying various plasmids with the relevant genes cloned downstream of tacOP were pulse labeled with H3[32P]O4 in the presence or absence of IPTG as inducer, and the proteins were separated on SDS–PAGE. The resulting autoradiographs are shown in Figure 3. First we reproduced the EIIBgl-dependent BglG phosphorylation demonstrated previously (Amster-Choder et al., 1989; Amster-Choder and Wright, 1990; Schnetz and Rak, 1990). When compared with the untransformed strain (Figure 3A, lane 1), a signal corresponding to the BglG protein can be observed without IPTG (Figure 3A, lane 2), and this signal increased strongly when IPTG was added (Figure 3A, lane 3). As expected from the previous data, the presence of salicin as substrate for EIIBgl transport completely inhibited phosphorylation of BglG (Figure 3A, lane 4). However, when bglG was expressed alone, an IPTG-dependent signal appeared at the position of BglG (Figure 3A, lanes 5 and 6). It disappeared when the same experiment was carried out with the pts deletion strain (Figure 3A, lanes 8 and 9). To unambiguously verify that the signal detected in the absence of EIIBgl is due to the phosphorylated BglG protein, we repeated the experiments in the wild type strain with a galK–bglG fusion which expresses a protein with an increased molecular weight and for which EIIBgl-dependent phosphorylation has previously been shown (Schnetz and Rak, 1990). Phosphorylated species were indeed observed at the expected positions and in an IPTG-dependent manner, not only in the presence of EIIBgl (Figure 3A, lanes 12 and 13) but also in its absence (Figure 3A, lanes 10 and 11). Again, no phosphorylation was observed in the presence of EIIBgl when salicin was present (Figure 3A, lane 14). As a control for BglG expression levels we performed [35S]methionine pulse–chase labeling and SDS–PAGE with the identical transformants used in Figure 3A and in the presence or absence of IPTG and salicin as given in the figure. The resulting autoradiograph revealed that neither coexpression of EIIBgl together with BglG nor the strain employed nor the presence of salicin had any appreciable effect on the amount or stability of BglG produced (data not shown), thus indicating that differences in 32P-signal strengths of BglG at identical IPTG concentrations reflect differences in the efficiency of BglG phosphorylation rather than differences in the amount of protein present. We also learned from this series of experiments that care had to be taken in adjusting the expression level of BglG. A 5-fold higher expression (as revealed from the [35S]methionine labeling; data not shown) resulted in an almost complete loss of the HPr-dependent phosphorylation signal of BglG (Figure 3A, lane 7). This effect could be attributed to the formation of insoluble inclusion bodies (Krieg, 1993; Chen et al., 1997a) which may sequester most of the available BglG protein. It may also explain why EIIBgl-independent phosphorylation previously remained undetected (Schnetz and Rak, 1990). Figure 3.Phosphorylation of BglG in vivo. (A) PTS-dependent phosphorylation of BglG also occurs in the absence of EIIBgl. Transformants of strains R1279 (pts+) and R1653 (Δpts) were grown with or without IPTG (0.1 mM) as inducer for tacOP driven expression of plasmid-encoded genes. Where indicated, salicin as substrate for EIIBgl was added. Lac repressor was provided from a second plasmid (pFDY226) with the exception of lane 7 (marked with an asterisk), where the lacI gene was present on the bglG-containing plasmid. In this case the plasmid had an increased copy number and 1 mM IPTG was added as inducer. Aliquots of cultures were labeled with H3[32P]O4, proteins separated by SDS–PAGE and gels subsequently exposed to X-ray films. Lane 1, R1279 (empty strain); lanes 2–4, R1279/pFDX3102 (bglG, bglF); lanes 5 and 6, R1279/pFDX2942 (bglG); lane 7, R1279/pFDY45 (bglG); lane 8, R1653/pFDX2942 (bglG); lane 9, R1653 (empty strain); lanes 10 and 11, R1279/pFDX3225 (galKΦbglG); lanes 12–14, R1279/pFDX3226 (galKΦbglG, bglF). (B) Massive PTS-sugar transport completely prevents phosphorylation of BglG in the absence as well as in the presence of EIIBgl. Transformants of strain R1279 were labeled as described in (A). In contrast to (A), gene bglF controlled by tacOP was provided by a lacI-containing plasmid (pFDX3283) in trans. IPTG as inducer for gene expression and sugars salicin, N-acetyl-D-glucosamine or mannitol as substrates for their respective EIIs were added where indicated. Lane 1, R1279 (empty strain); lanes 2 and 3, R1279/pFDX2942 (bglG)/pFDY226; lanes 4 and 5, R1279/pFDX3293 (bglG, nagE)/pFDY226; lanes 6 and 7, R1279/pFDX3295 (bglG, mtlA)/pFDY226; lanes 8–10, R1279/pFDX2942 (bglG)/pFDX3283 (bglF); lanes 11 and 12, R1279/ pFDX3293 (bglG, nagE)/pFDX3283 (bglF); lanes 13 and 14, R1279/pFDX3295 (bglG, mtlA)/pFDX3283 (bglF). Download figure Download PowerPoint Underphosphorylation of the PTS leads to dephosphorylation and concomitant inactivation of BglG According to previous models (see Introduction) EIIBgl, in the absence of β-glucosidic substrates, inhibits BglG by phosphorylation and reactivates it by dephosphorylation when substrates become available (negative-feedback control). The data presented above extend this model and demonstrate that activity of BglG is additionally under positive control of the PTS, dependent upon HPr-catalyzed phosphorylation, which should occur at a different site than that catalyzed by EIIBgl. However, the observation that the presence of salicin as sugar substrate completely prevents phosphorylation of BglG (Figure 3A, lanes 4 and 14) at first seemed to contradict this conclusion. One has to bear in mind, however, that analysis of BglG activity (Figure 2) required single-copy situations, while the salicin-dependent dephosphorylation of BglG, which was performed under in vivo conditions (Figure 3A), could only be demonstrated when bglG and bglF were overexpressed from a plasmid. We therefore reasoned that the HPr-catalyzed phosphorylation of BglG was prevented due to massive salicin transport catalyzed by unphysiologically high concentrations of EIIBgl. This might drain the entire phosphorylation capacity of the PTS towards the incoming sugar. To investigate this possibility, we performed two types of analysis. We analyzed phosphorylation of BglG expressed from a plasmid (Figure 3B) and in parallel studied the activity of BglG expressed from a single-copy gene, in both cases in the presence of overproduced EIIs other than EIIBgl (Figure 4). To indicate the position of the phosphorylated form of BglG and IPTG dependence of its expression, the phosphorylation assays shown in Figure 3A, lanes 1, 5 and 6, were repeated in Figure 3B, lanes 1, 2 and 3, respectively. Neither overexpression of nagE (encoding EIINag) nor that of mtlA (encoding EIIMtl) led to significant alterations in the state of phosphorylation of BglG (Figure 3B, lanes 4 and 6, respectively). The presence of the respective sugar substrates, N-acetyl-D-glucosamine (Nag) and D-mannitol (Mtl), however, completely prevented phosphorylation of BglG (Figure 3B, lanes 5 and 7). Lanes 8–10 correspond to lanes 2–4 in Figure 3A, with the difference that EIIBgl was delivered from a second plasmid in trans. Again, the phosphorylated form of BglG appeared in an IPTG-dependent manner (Figure 3B, lanes 8 and 9) and disappeared when salicin was present (lane 10). As expected, this pattern did not alter, when in addition, nagE (lane 11) or mtlA (lane 13) was expressed as an operon fusion with bglG. However, the presence of the respective sugar substrates (Nag; Figure 3B, lane 12 or Mtl; lane 14) almost completely prevented phosphorylation of BglG. Again, we verified by [35S]methionine pulse–chase labeling and SDS–PAGE with the identical transformants that the conditions employed had no unexpected effects on the expression level of BglG (data not shown). We can thus conclude that massive transport of a PTS substrate generally causes disappearance of the phosphorylated form of BglG, regardless of the presence of EIIBgl. Figure 4.Massive PTS sugar transport drastically decreases BglG activity. Strains R1752 and R1958 carrying tacOP-controlled genes bglG–bglF and bglG, respectively, integrated into the chromosome were cotransformed with antitermination reporter plasmid pFDX2676 and one of the following plasmids for tacOP-controlled expression of different EIIs (where indicated): pFDX3297 (nagE; EIINag), pFDX3298 (mtlA; EIIMtl), pFDX3299 (bglF; EIIBgl). Given are β-galactosidase enzyme activities synthesized by these transformants in presence or absence of IPTG as inducer for gene expression and the various PTS sugar transport substrates. −, not determined. Download figure Download PowerPoint To correlate the phosphorylation data with the activity of BglG, we performed the antitermination assay with the transformants used in Figure 2, columns 1. These cells were transformed with additional plasmids carrying nagE, mtlA or bglF under tacOP control (Figure 4). For comparison, the values without an additional plasmid (taken from Figure 2, column 1) are also included. It can be seen that the additional expression of nagE in the absence of N-actetyl-D-glucosamine (Nag) as substrate, or mtlA in the absence of D-mannitol (Mtl) did not influence the BglG activity originating from the bglG gene or the bglG–bglF operon. In contrast, the presence of the respective sugar substrates reduced activity to nearly background, not only when the bglG gene was expressed alone, but also when the bglG–bglF operon was expressed and salicin was included as inducer. These results support the notion derived from the phosphorylation data that phosphoryl groups within the PTS may be limiting for the activation of BglG. They support our conclusion that BglG must be phosphorylated in order to be active. Further support is derived from the data in the last column in Figure 4:overexpression of bglF from a plasmid in the strain containing bglG alone led to a drastic reduction of BglG activity even in the presence of salicin, and reduced it to almost background levels in the strain carrying the bglG–bglF operon. Does the limitation of phosphoryl group availability play a role under physiological conditions? We wanted to learn whether the inhibitory effect on BglG activation seen under conditions of overproduction of EIIs and in the presence of the respective sugar substrates points to a limited pool of phosphoryl groups under more physiological conditions. From this we hoped to learn more about its possible role in keeping the activity of BglG low under natural conditions when other, perhaps more favorable sugars are available. To this end we performed the antitermination assay as in Figure 2, columns 1 but replaced glycerol with other sugars. In the case of the bglG–bglF operon, salicin was in addition present as substrate for EIIBgl and thus as inducer of BglG activity. The results are given in Figure 5. Glycerol is a non-PTS sugar, while the other sugars used are all substrates of the PTS. Compared with glycerol, all PTS sugars indeed led to a reduction of BglG activity to different extents. Figure 5.Utilization of PTS sugars reduces BglG activity to different degrees. Strains R1958 (light gray histograms) and R1752 (dark gray histograms) carrying the chromosomally integrated tacOP driven bglG and bglG–bglF expression cassettes, respectively, were transformed with antitermination reporter plasmid pFDX2676. Cells were grown in M9-medium with the indicated sugar as carbon source. IPTG was added to induce synthesis of BglG and BglG together with EIIBgl, respectively. In the case of R1752 salicin was additionally present to release BglG from negative control exerted by EIIBgl. Download figure Download PowerPoint Discussion In view of the previous finding that the PTS negatively controls activity of BglG by EIIBgl-catalyzed phosphorylation it might have been expected that any mutation preventing this phosphoryl group transfer would result in unrestrictedly high activity of BglG. Indeed, activity of BglG was high and constitutive in the absence of EIIBgl (Figures 1 and 2) or when cysteine residue 24 within EIIBgl, the phosphoryl group donor site for BglG phosphorylation, was replaced by a serine (Krieg, 1993; Chen et al., 1997a). However, contrary to expectation, BglG turned out to be inactive when the central energy-coupling proteins of the PTS, EI and HPr, were absent (Figure 2), suggesting a second level of BglG activity control requiring the PTS. Since evidence has been presented that the active form of BglG is a dimer (Amster-Choder and Wright, 1992), the positively acting signal may be necessary for dimer formation. This requirement can, however, be overcome, at least in part, by an increase in its concentration above the physiological level (Figure 1). Similar observations have been reported for LicT of the BglG family of antiterminators (Krüger and Hecker, 1995; Krüger et al., 1996) and the related transcriptional activator LevR (Stülke et al., 1995). Detailed genetic complementation analyses revealed that phospho-HPr is required for BglG activity, and it turned out that for this activation, HPr can be substituted by DTP, a component of the fructose-inducible PTS encoded by the fru operon. Suppression of HPr deficiency by DTP has been observed before for PTS transport functions, either in mutants lacking the fru repressor (Saier and Ramseier, 1996) or in cultures induced with fructose (Saier et al., 1970). The possible physiological role of the cross-talk observed here for BglG activation remains to be determined. Since our results suggested that BglG might be directly phosphorylated by HPr, we performed in vivo phosphorylation analyses. Indeed, phosphorylated BglG species were detected in the absence of EIIBgl, under conditions where BglG activity was constitutively high (Figure 3). On the other hand, BglG was inactive and not
Referência(s)