![[Feedback]](/icons/banner/feedbackACT.gif)
![[For Subscribers]](/icons/banner/subscriberACT.gif)
![[Archive]](/icons/banner/archiveACT.gif)
![[Search]](/icons/banner/searchACT.gif)
![[Contents]](/icons/banner/tocACT.gif)
Microbiology and Molecular Biology Reviews, September 2004, p. 474-500, Vol. 68, No. 3
1092-2172/04/$08.00+0 DOI: 10.1128/MMBR.68.3.474-500.2004
Copyright © 2004, American Society for Microbiology. All Rights Reserved.
Bacterial Transcriptional Regulators for Degradation Pathways of Aromatic Compounds
David Tropel1 and Jan Roelof van der Meer1,2*
Swiss Federal Institute for Environmental Science and Technology (EAWAG), Dübendorf,1
Department of Fundamental Microbiology, University of Lausanne, Lausanne, Switzerland2
Human activities have resulted in the release and introduction into the environment of a plethora of aromatic chemicals. The interest in discovering how bacteria are dealing with hazardous environmental pollutants has driven a large research community and has resulted in important biochemical, genetic, and physiological knowledge about the degradation capacities of microorganisms and their application in bioremediation, green chemistry, or production of pharmacy synthons. In addition, regulation of catabolic pathway expression has attracted the interest of numerous different groups, and several catabolic pathway regulators have been exemplary for understanding transcription control mechanisms. More recently, information about regulatory systems has been used to construct whole-cell living bioreporters that are used to measure the quality of the aqueous, soil, and air environment. The topic of biodegradation is relatively coherent, and this review presents a coherent overview of the regulatory systems involved in the transcriptional control of catabolic pathways. This review summarizes the different regulatory systems involved in biodegradation pathways of aromatic compounds linking them to other known protein families. Specific attention has been paid to describing the genetic organization of the regulatory genes, promoters, and target operon(s) and to discussing present knowledge about signaling molecules, DNA binding properties, and operator characteristics, and evidence from regulatory mutants. For each regulator family, this information is combined with recently obtained protein structural information to arrive at a possible mechanism of transcription activation. This demonstrates the diversity of control mechanisms existing in catabolic pathways.
Human activities have resulted in the release and introduction into the environment of a plethora of aromatic chemicals. Although the input of these synthetic chemicals may be smaller than the total amount of aromatic compounds released from decaying plant material, their novel structures or their quantities as single pure molecules induce major changes in microbial communities, which are the major recyclers of organic chemicals in nature (76). For example, large quantities of a single organic solvent offer a very specific metabolizable carbon source for specialized groups of microorganisms (161) and thus can act selectively for enrichment of those groups in the environment (87, 154). Compounds such as chlorinated solvents, herbicides, and pesticides often carry uncommon chemical structures, side chains, or functional groups, which can have toxic effects or provide new carbon sources for bacteria which have adapted their metabolism to degrade the compounds (191, 207, 261). The interest in discovering how bacteria are dealing with hazardous environmental pollutants has driven a large research community and has resulted in important biochemical, genetic, and physiological knowledge about the degradation capacities of microorganisms (40, 43, 76, 146, 191, 206, 207). A large variety of metabolic pathways have been discovered in very different microorganisms, fueling up-to-date online databases such as the Biocatalysis/Biodegradation Database (57). Knowledge about the biodegradation capacities of microorganisms is being applied directly in bioremediation practice (240), and individual biotransformation reactions are potentially very interesting for incorporation into chemical synthesis (230).
Less obviously associated with bioremediation applications, green chemistry, or production of pharmacy synthons have been studies of the regulatory mechanisms, which govern the expression of specific catabolic pathways (105). However, regulation of catabolic pathway expression has attracted the interest of numerous different groups, who have tried to unravel the molecular partners in the regulatory process, the signals triggering pathway expression, and the exact mechanisms of activation and repression. More recently, information about regulatory systems has attracted interest, with the potential of these systems being used as sensory mechanisms in whole-cell living bioreporters, genetically modified bacteria which can be used as sensors to measure the quality of aqueous, soil, and air environments (49, 96, 242). It soon was discovered that a large diversity of regulatory systems existed for mediating the expression of catabolic pathways. Furthermore, many related catabolic pathways did not carry the same regulatory system, which led to the hypothesis that regulatory systems and their target operons do not necessarily coevolve but seem to become associated independently (23, 39). As genomic, genetic, and biochemical data accumulated, regulatory proteins for catabolic gene expression were classified into other protein families, showing that they (as expected) were not unique to typical biodegradation pathways per se. In fact, many aromatic degradation pathways, typically those involved in the metabolism of aromatic amino acids, hydroxylated benzoates and phenylpropionic acids, ubiquinones, and aromatic amines, are very widespread among microbial species (43), leaving perhaps only the catabolic pathways for toxic and xenobiotic compounds to the groups of microorganisms usually considered for biodegradation (i.e., pseudomonads, sphingomonads, Rhodococcus spp., Ralstonia spp., and Burkholderia spp.). One would thus have to conclude that the capabilities of regulatory proteins to react specifically to substrates or intermediates of catabolic pathways have evolved many times independently; this is explained in more detail below.
Although excellent reviews exist on the diversity of catabolic pathways (76, 206), on evolutionary aspects (99, 262, 271), and on different details of specific regulator families (71, 139, 224, 233), no concise overview of the different regulatory systems involved in biodegradation pathways (or catabolic pathways, as we will refer to them often) exists in the literature, although a shorter treatise was published by Diaz and Priets (44). In our opinion, the topic of biodegradation is relatively coherent, and thus treatment of the different regulatory systems involved in catabolic pathways also deserves a coherent overview—although it is clear that there are no special "catabolic" regulatory proteins. In addition, several three-dimensional structures of regulatory proteins have recently been resolved, which allows a much more detailed prediction of how the actual activation and repression mechanisms of regulatory proteins take place. This might solve many questions which have arisen from more traditional genetic and biochemical approaches to transcription activation and repression. The current view of regulatory proteins is that of a set of sophisticated protein-DNA-RNA polymerase interaction machineries with sufficient diversity to promote or inhibit DNA transcription by RNA polymerase. What remains largely elusive, however, is how the machinery is turning.
The structure of this review is relatively straightforward. The various known regulatory families are treated individually, as far as possible with examples of catabolic pathways. Specific attention has been given to describing the genetic organization of the regulatory genes, promoters, and target operon(s); this is followed by a discussion integrating the knowledge of signaling molecules, DNA binding properties and operator characteristics, regulator-DNA structure, and evidence from regulatory mutants to arrive at a possible mechanism of transcription activation. What is not specifically treated in this review is how regulatory systems themselves are embedded in the host's physiology and controlled by other global regulatory systems. For information about these aspects, readers are referred to other specific overviews, such as references 23, 232, and 265. For sake of the overview, we have refrained from too many in-depth details of very specific individual transcription regulators.
|
LysR FAMILY OF TRANSCRIPTIONAL REGULATORS
|
|---|
Catabolic Operons Controlled by LysR-Type Regulators
LysR-type transcriptional regulators (LTTRs) comprise the largest family of prokaryotic regulatory proteins identified so far (84, 224). The family has expanded to over 100 members that have been identified in diverse bacterial genera. This diversity is also reflected in LTTRs associated with degradation pathways of aromatic compounds (Table 1) (44). A large group of LTTRs regulates a single target operon only, such as CatR controlling catBCA expression for catechol metabolism in Pseudomonas putida (215), ClcR controlling the clcABDE operon of plasmid pAC27 (33), TcbR controlling the tcbCDEF operon in plasmid pP51 of Pseudomonas sp. strain P51 (263), and CbnR controlling the cbnABCD operon for chlorocatechol metabolism in Ralstonia eutropha (163). TfdR (and/or its identical twin TfdS) also regulate an operon coding for chlorocatechol metabolism (i.e., the tfdDCEFB genes), but the same regulators also control the expression of two other operons, tfdA and tfdDIICIIEIIFIIBIIK, all of which are involved in degradation of 2,4-dichlorophenoxyacetate in R. eutropha JMP134 (and are located on plasmid pJP4 [Fig. 1A ]) (127). Similarly, CatR of P. putida can coregulate the expression of the pheBA operon for phenol degradation and of the catBCA genes when this operon is provided on an additional plasmid (104). Like TfdR and TfdS, the NahR protein is a master regulator for the regulon of naphthalene degradation and acts by controlling expression from both the nah operon, required for the metabolism of naphthalene to salicylate and pyruvate, and the sal operon, encoding the enzymes for salicylate conversion (225, 228, 274). In this case, both operons are present on the NAH7 plasmid. Two LTTRs have been implicated in the transcriptional control of benzoate degradation in Acinetobacter sp. strain ADP1. In this microbe, BenM regulates expression from the benABCDE and benPK operons, which encode the enzymes for conversion of benzoate to catechol. CatM also regulates expression from benPK but in addition regulates expression from the catBCIJFD operon, leading to further metabolism of cis,cis-muconate to the tricarboxylic acid cycle (Fig. 1B). The catA gene, encoding catechol 1,2-dioxygenase, which converts catechol to cis,cis-muconate, is not present in the two operons but is also regulated by both CatM and BenM (34, 212). CatM and BenM also interfere with expression of the pobA gene for 4-hydroxybenzoate metabolism, perhaps by preventing the expression of the pcaK uptake system (16). Although LTTRs are often associated with ortho-cleavage pathways of catechol, they are also involved in a large variety of other degradation pathways (Table 1). For example, NtdR controls the genes encoding nitroarene dioxygenase in Acidovorax sp. strain JS42 (125); the TsaR protein controls the expression of the tsaMBCD genes, which encode the first steps in the degradation of p-toluenesulfonate in Comamonas testosteroni T-2 (253); and BphR2 is involved in the regulation of a biphenyl catabolic gene cluster of Pseudomonas pseudoalcaligenes KF707 (268). Preliminary data for regulation of the biphenyl pathway indicate that BphR2 is activating the expression of a first operon [bphA1A2(orf3)A3A4] whereas a second regulatory gene (bphR1, a regulator of the GntR family) is activating the expression of the second operon (bphX1X2X3D). LTTRs have also been identified in 
-hexachlorocyclohexane degradation in Sphingomonas paucimobilis (148), in aniline degradation in P. putida (67), and in catabolism of 3-phenylpropionic acid in Escherichia coli K-12 (42). A longer list of LTTRs involved in transcriptional control of aromatic compound degradation is presented in reference 44.
View this table:
[in this window]
[in a new window]
|
TABLE 1. Regulatory proteins from the LysR family involved in expression control of pathways for degradation of aromatic compounds
|
|
View larger version (31K):
[in this window]
[in a new window]
|
FIG. 1. Typical examples of the genetic organization of pathways for degradation of aromatic compounds. The regulatory genes are indicated in grey. (A) The tfd operons of R. eutropha JMP134(pJP4). The LysR-type transcription regulator TfdR/S regulates the tfdA, tfdC, and tfdDII promoters. (B) The ben and cat operons of Acinetobacter sp. strain ADP1. benM and catM code for LysR-type transcription regulators, which act on four promoters: benP, benA, catA, and catB. (C) The pca and pob clusters of Acinetobacter sp. strain ADP1. The genes pcaU and pobR code for IclR-type regulators, acting on the pcaI and pobA promoters. (D) The hpa cluster of E. coli with HpaA as the XylS/AraC-type regulator. (E) The xyl operons for toluene degradation present on the TOL plasmid of P. putida mt2. XylR is the NtrC-type regulator acting on Pu and Ps, whereas XylS is the exemplar for the XylS/AraC family and acts on the Pm promoter. (F) The paa cluster in E. coli strain W, with the GntR-type regulator PaaX regulating the paaA promoter. (G) The aph cluster in C. testosteroni strain TA441, encoding the aphS (GntR-type) and aphR (XylR-type) regulators. (H) The cym and cmt clusters in P. putida F1 with the TetR-type regulator cymR. (I) The tod cluster of P. putida DOT-T1 with the two-component regulatory system todST. Straight arrows represent transcripts produced after specific regulatory action of the indicated regulator. Parabolic arrows point to the site of action of a specific regulator in cases when multiple regulators are involved. Hooked arrows indicate specific promoter names, where necessary.
|
|
In general, the gene for LTTRs lies upstream of its target-regulated gene cluster and is transcribed in the opposite direction (Fig. 1A and B), but exceptions to this rule exist. For example, the gene for the SalR regulator in Acinetobacter sp. strain ADP1 is present within the transcriptional unit salRA. The salA gene is coding for salicylate hydroxylase, which converts salicylate to catechol. Experimental evidence suggests that SalR is regulating the expression of itself and of salA (101). The phnS gene (coding for a further uncharacterized LTTR) is cotranscribed as the first gene of an operon including the phnFECDAL catabolic genes for naphthalene and phenanthrene degradation in Burkholderia sp. strain RP007 (119).
All identified LTTRs involved in aromatic degradation pathways act as transcriptional activators for their target metabolic operons in the presence of a chemical inducer, which is usually a pathway intermediate (33, 128, 212, 215, 227). In some cases, the effectors are (substituted) muconates which have lost their aromatic character, whereas in other systems aromatic compounds act as effectors (e.g., salicylate, catechol, and nitrotoluene [Table 1]). There is also experimental evidence for competition of several compounds on the regulatory protein. For example, fumarate reversibly inhibits the formation of the clcA transcript in in vitro transcription assays in the presence of purified ClcR and 2-chloro-cis,cis-muconate (138). The presence of cis,cis-muconate decreases the affinity of the BenM protein for benzoate (31). All LTTRs repress their own expression, and both autorepression and activation of the catabolic operon promoter are exerted from the same binding site, which is called the regulator or repressor binding site (RBS) (20, 216, 227). Autorepression was not influenced by the presence of an inducer in the case of BenM (20), but expression of the clcR and tcbR promoters was slightly enhanced in the presence of an inducer (33, 263). Relatively few data exist on autorepression mechanisms, since most studies on LTTRs have focused on the mechanisms of target gene activation.
Structure and Conformation
LTTRs involved in the degradation of aromatic compounds are composed of 394 to 403 amino acid residues with a molecular mass of between 32 and 37 kDa. All the evidence presented so far points to tetramers being the active form of LTTRs. ClcR and CatR have been identified as dimers in solution (32, 181), but two dimers are needed to bind DNA (141). CatM, BenM, and CbnR were found to form tetramers (20, 30, 155), and NahR probably also acts as a tetrameric form (226). The well-studied LTTR CysB, which is not involved in the activation of pathways for aromatic-compound degradation, also forms tetramers in solution (88, 145). Only TsaR remains as monomer in solution (253). LTTRs have a conserved domain organization, which has been determined from mutagenesis studies and sequence alignments (224). A DNA binding region with a predicted helix-turn-helix (HTH) motif is located in the 66 N-terminal amino acid residues of the protein, two regions located between residues 95 to 173 and residues 196 to 206 are involved in inducer recognition, and one region between residues 227 and 253 is supposed to be involved in multimerization (Fig. 2A).
View larger version (25K):
[in this window]
[in a new window]
|
FIG. 2. Schematic domain organization of the different regulator families. The domain containing the HTH DNA binding motif is indicated in black. The domain bound by the chemical inducer is indicated in white.
|
|
Recently, the first complete LTTR has been crystallized (CbnR) (155). Crystals of BenM and CatM devoid of their N-terminal DNA binding domain have also been obtained, but their structures have not been resolved yet (29). CbnR crystallized as a tetramer (Fig. 3) with two main parts, the four DNA binding domains (residues 1 to 58) and a central body (155). The four DNA binding domains have no interactions with each other, whereas the central body of the tetramer is composed of four intertwined regulatory domains (residues 88 to 294 of each subunit). The tetramer structure can be regarded as a dimer of a dimer, whereby each dimer is composed of two subunits in different configurations. The two subunits in each dimer are connected through the coiled-coil linker (residues 59 to 87), which at the same time separates the DNA binding domain and the central body (Fig. 3). Subunit A of the AB dimer interacts with the regulatory domains of both subunits Q and P of the other (PQ) dimer, whereas subunit B interacts only with P (Fig. 3). The structure seems to allow easy transmission of conformational changes, for example in the central body, the DNA binding domains. The DNA binding motifs in the tetramer are presented in such a way as to form a V-shaped structure, which matches exactly with the distance and configuration of the two DNA binding sites (Fig. 3). Although the mode of action and the location of the effector binding pockets still need better definition, this structural model beautifully fits and explains previous experimental evidence with the LTTRs.
View larger version (43K):
[in this window]
[in a new window]
|
FIG. 3. Tetrameric structure of CbnR in side (A) and top (B) views (155). Subunits A, B, P, and Q are shown in yellow, magenta, cyan, and green, respectively. The symmetry axis in the molecule is shown as a red arrow. The location of the coiled-coil linkers (see the text) and DNA binding domains (DBD) are indicated, which would impose a 60°C bending angle on the binding site on the DNA. Reprinted from reference 155 with permission from the publisher and from the authors.
|
|
Mechanisms of Activation
Due to their tetrameric form, LTTRs interact with several sites on the DNA of the promoter region. Classically, DNA interactions are shown by DNase I footprinting techniques and purified regulatory protein. Interestingly, DNase I cleavage patterns of the protein-bound nucleotide regions were similar for ClcR, CatR, TcbR, CbnR, and, to a lesser extent, NahR. The different interactions are explained in more detail for ClcR. ClcR binds the clcA promoter irrespective of the presence of inducer (32, 141, 182). However, in the absence of inducer, ClcR protects a 27-bp region (RBS) from –79 to –53 and a 10-bp region (activator binding site [ABS]) from –37 to –28 relative to the transcription start site. Each region is supposed to be bound by one of the two dimers in the tetramer (Fig. 4). The RBS contains an interrupted inverted repeat, ATAC-N7-GTAT, with the consensus LTTR binding motif T-N11-A (224). Two hypersensitive bands (–52/–51 and –42) also show up on ClcR-bound promoter DNaseI footprints. This may reflect the bending imposed on the DNA by the V-shaped configuration of the tetramer (Fig. 3). In the presence of inducer, the hypersensitive band at –42 disappeared and occupation of the ABS shifted from –37 to –41 (141), suggesting that the bending angle is relaxed on interaction with the effector. Conformational changes on effector binding could be detected for benzoate and cis,cis-muconate binding to BenM (31). Measurements of the bending angle of the clcA promoter in the absence (71°) and in the presence (55°) of effector corroborate this idea of bending relaxation (140). Similar bending angles were reported for CbnR at the cbnA promoter in the absence and presence of inducer, although for CbnR no changes in DNase I footprints were observed as for ClcR (163). The role of the RBS and ABS is not completely clarified. Although ClcR and CbnR contacted both the RBS and ABS in the absence of inducer, CatR did not (215, 216). It was concluded that the regulatory proteins have a higher binding affinity to the RBS than to the ABS, since a fragment with only the ABS is not bound by ClcR or CatR (139, 181). However, contacts to the ABS are supposed to be necessary for mediating interactions with RNA polymerase (RNAP), and fragments with only the RBS do not lead to in vitro transcript production in the presence of ClcR, RNAP, and inducer (139).
View larger version (21K):
[in this window]
[in a new window]
|
FIG. 4. Schematic steps in transcription activation by LysR-type transcription regulators. (A) Two subunits of the regulator tetramer binds to the RBS and two other subunits bind to the ABS present on the –35 promoter region. (B and C) In the presence of an inducer, the regulatory protein shifts its interaction on the ABS to –42. By interacting with the  -CTD domain of the RNAP, it directs this to bind a so-called UP-DNA sequence motif, which is located between the RBS and ABS (B). This increases the binding affinity of RNAP for the promoter and initiates transcription (C). mRNA is drawn as a zig-zagged arrow.
|
|
Further contacts seem to occur between the regulatory protein complex and the C-terminal domain of the -subunit (-CTD) of RNAP. This was concluded from the lack of mRNA synthesis in in vitro transcription experiments with CatR (or ClcR), the catB (clcA) promoter, cis,cis-muconate, and RNAP devoid of the -CTD (28, 139). A direct interaction of NahR with the -CTD of RNAP has been demonstrated by using a yeast two-hybrid system (178), and this interaction was not influenced by the presence of salicylate. Chugani et al. (28) postulated that the regulatory complex was directing the -CTD of RNAP to a region between the RBS and the ABS, which they called the UP-motif (Fig. 4). Interaction with the UP-motif would increase the affinity of RNAP to the promoter (28). However, exactly how this process would lead to transcription activation is not known. The degree of DNA supercoiling has also been implicated in controlling the level of transcription activation from LTTR-dependent promoters. This has been concluded from in vitro transcription experiments with CatR, in which supercoiled templates produced no mRNA transcript except in the presence of cis,cis-muconate whereas linearized templates did (28). Perhaps the regulatory protein can overcome the torsional constraint on the DNA in the presence of inducer, thereby facilitating open-complex formation by RNAP.
In some cases, a third DNA binding site further downstream of the transcriptional start site has been identified. For example, CatR binds with low affinity to a region between +162 and +196 of the catB gene (27). This site has been named the inhibitor binding site (IBS) and is supposed to prevent excessively high expression from the catBCA operon by titrating CatR protein. The affinity of binding of CatR to the IBS increased at higher inducer concentrations, and catB promoter fragments containing the IBS had a three- to fourfold lower expression than those without IBS (27). IBS regions have also been identified in the pheBA operon and clcABD operon (27).
BenM has the unique peculiarity (for the moment) that the inducers benzoate and cis,cis-muconate can have synergistic effects on the activation process compared to the effects of each inducer alone (20, 31, 34). In addition, BenM represses benA transcription in the absence of inducer. Benzoate and cis,cis-muconate alone had very different effects on the patterns observed in DNase I footprints of BenM on the benA promoter than did both inducers simultaneously. In the absence of inducer, BenM bound two areas with dyad symmetry, site 1 (ATAC-N7-GTAT) at positions –57 to –71 and site 3 (ATTC-N7-GTAT) at positions –5 to –19. Several hypersensitive sites were detected (at positions –50, –45, –39, –36, –34, –29, and –24), suggesting again a clear bending of the DNA imposed by the BenM regulatory complex. In the presence of cis,cis-muconate or benzoate, BenM still protected site 1 but no longer protected site 3. When both inducers were present simultaneously, the number of hypersensitive sites was reduced and instead a region called site 2 (which also contained a dyad symmetry motif, ATAC-N7-GTGT, located between –36 and –50) was protected from DNase I cleavage (20). The authors hypothesized that BenM binding to site 3 (which overlaps with the –10 region) causes the observed repressive effect in the absence of inducer. In the presence of inducer, is BenM released from site 3, enabling RNAP to access the promoter, whereas BenM binding to site 2 would lead to more productive contacts to RNAP and a higher transcriptional output.
|
IclR FAMILY OF TRANSCRIPTIONAL REGULATORS
|
|---|
Catabolic Operons Controlled by IclR Regulators
IclR-type regulators have a similar structure as the LysR-type regulators (224), but rather dissimilar amino acid sequences distinguish the two families. IclR-type regulators are generally transcriptional repressors (86, 132, 156, 245); however, those which control catabolic pathways have all been described as activators (Table 2). For example, PcaU of Acinetobacter sp. strain ADP1 (75), PcaR of P. putida (213), PcaR of Agrobacterium tumefaciens, and CatR and PcaR of Rhodococcus opacus 1CP (58, 59) are activators for the ortho-cleavage pathways which they regulate. However, in the absence of inducer they may still act as repressors, as was shown for PcaU (254). Further IclR members are MhpR, which activates the meta-cleavage pathway in 3-(3-hydroxyphenyl)propionic acid degradation by E. coli (61); PobR, which is the activator for the 4-hydroxybenzoate degradation pathway in Acinetobacter sp. strain ADP1 (45), and OhbR, controlling the genes for the oxygenolytic ortho dehalogenation of halobenzoates (256).
In general, the gene for the IclR-type regulator lies upstream of its target gene cluster and is transcribed in the opposite direction (Fig. 1C). An exception is the pcaR gene of P. putida (160). PcaR actually controls four distinct gene clusters required for the degradation of protocatechuate to tricarboxylic acid cycle intermediates and of the 4-hydroxybenzoate transporter PcaK (160). The pcaR gene lies upstream of pcaK and is transcribed in the same direction (90, 159, 213). PcaR from P. putida and PobR and PcaU from Acinetobacter sp. strain ADP1 repress their own expression (46, 81, 254). It is thought that the mechanism of autorepression is different among IclR-type regulators, since not all of them bind at the same position on the promoter region (75, 81, 213) and since for some of them the addition of effectors changes the expression of the regulatory gene itself. For example, expression of pcaR and pobR is independent of the usual effectors for PcaR and PobR (46, 81) but pcaU expression increases in the presence of effectors for PcaU (75).
Structure and Conformation
The size of IclR-type regulators is around 238 to 280 amino acid residues (25 to 30 kDa) (59, 61, 82, 112, 256). IclR-family members have an HTH DNA binding motif in the N-terminal domain (45, 61) and a C-terminal domain involved in subunit multimerization and effector binding (112). This was confirmed by the crystal structure of IclR from Thermotoga maritima 0065 (Fig. 5) (275), which showed that the amino acid residues previously identified to be involved in PobR inducer specificity (112) are present in IclR and comprise a pocket in the C-terminal domain (275). Although PcaU and PcaR from P. putida formed dimers in solution (81, 193), IclR crystallized as a dimer of a dimer with an asymmetric configuration (275). The two subunits within one IclR dimer interact solely at the interface of their DNA binding domains. As a consequence, the distance between the HTH motifs within one dimer is relatively short and results in a structure favorable for binding relatively short (12- to 14-bp) palindromic DNA sequences with specific contacts predominantly in the major groove of the DNA. The C-terminal domains do not contact each other in the dimer, but they do bridge with the C-terminal domains from the neighboring dimer, which is oriented in asymmetric fashion (Fig. 5). It is not really clear how the tetramer interacts with the DNA. Mass spectrometric data revealed that four IclR subunits are present per DNA containing one single palindrome (see below). The presumed ligand binding region is close to the region involved in tetramerization, suggesting that ligand binding and tetramerization may be linked. Moreover, the tetramer seems to be the active DNA binding form since, in mass spectrometric studies, four IclR subunits bound one synthetic DNA containing one single palindrome. The stoichiometry of two protein subunits for one DNA molecule was never detected (48).
View larger version (42K):
[in this window]
[in a new window]
|
FIG. 5. IclR dimer and tetramer arrangements as derived from the crystal structure (275). (A) The dimer interface is formed exclusively between the two HTH DNA binding domains. Monomers are colored red or green. (B) The tetramer viewed from the top is composed of two asymmetric dimers shown in four different colors: red and green for one asymmetric unit, and yellow and magenta for the other. The tetramer interface is formed exclusively between signal binding domains. Reprinted from reference 275 with permission from the publisher and from the authors.
|
|
Mechanism of Transcription Activation
There is no clear consensus on the binding site for IclR members (Table 2). For example, the MhpR binding site is formed by one 15-bp palindrome that lies 50 to 66 nucleotides upstream of the mhpA transcription start site (252), but those of PcaU and PobR are three perfect 10-bp sequence repetitions that lie between 53 and 94 bp upstream of the pobA and pcaI transcription start sites, respectively (46, 75, 193). Two of the 10-bp sequence repetitions form one palindrome, and the third repeat is oriented opposite to the second one but separated by another 10 bp (193). In contrast to those, the PcaR binding site is formed by a series of 15 nucleotides present twice in the pcaI promoter and overlapping with the –10/–35 promoter region (81). One might thus conclude that when the regulators conform to their binding sites, they must be different between MhpR, PobR, PcaU, and PcaR.
IclR-type regulators bind their promoter DNA in the absence of effector, and adding effector molecules had no effect on the affinity of the protein-DNA interactions displayed by purified PobR, PcaU, PcaR, and MhpR (46, 75, 81, 252). When, however, the chemical effector was added to a mixture of regulator (in this case PcaR), purified 
70-RNAP, and a pcaI promoter fragment, the formation of a PcaR-RNAP-DNA complex was enhanced compared to the situation without effector (81). The authors suggested that the role of the regulatory protein might be to favor the recruitment of RNAP to the promoter, perhaps by optimizing the critical distance between the –35 and –10 elements in the pcaI promoter from 16 to 17 bp (81). The same effect of enhanced complex formation was also found for SoxR and MerR (6, 85). Also, in the mhpA promoter the –35 and –10 elements are separated by 16 bp (252). However, the pobA promoter (regulated by PobR) already has a spacing of 17 bp (75), and thus the generality of this promoter distance optimization as an activation mechanism for IclR-type regulators is debatable.
Catabolic Operons Controlled by AraC/XylS-Type Regulators
For years the XylS protein was the only member of the AraC family which was involved in expression control of a catabolic operon, namely, the meta-cleavage pathway operon for degradation of meta-toluate located on the TOL plasmid in P. putida (92). Recent alignment studies have shown that more than 300 proteins, some of which may be involved in the control of catabolic pathways, contained a typical stretch in the C-terminal part of about 100 amino acids that would classify them as AraC/XylS-type regulators (251). The remaining part of the proteins can be very different, though (see below). AraC/XylS-type regulators for catabolic operons generally act as transcription activators in the presence of a chemical effector molecule (Table 3). Some of the more recently discovered catabolic gene regulators of the AraC/XylS family are PobC of P. putida WCS358 (13), PobR of Azotobacter chroococcum (198), and PobR of Pseudomonas sp. strain HR199 (173). PobC and PobR regulate the expression of a pobA gene for p-hydroxybenzoate hydroxylase. HpaA is a XylS-type protein found in E. coli and regulates the expression of the hpaCB genes for p-hydroxyphenylacetate hydroxylase (195). Other family members identified in degradation pathways are CbdS, controlling the 2-halobenzoate dioxygenase genes in Burkholderia sp. strain TH2 (246); BenR, controlling the benzoate 1,2-dioxygenase operon in P. putida (36); CadR, controlling the genes for 2,4-dichlorophenoxyacetic acid degradation in Bradyrhizobium sp. strain HW13 (109); and IpbR, which regulates the expression of isopropylbenzene degradation in P. putida RE204 (51).
View this table:
[in this window]
[in a new window]
|
TABLE 3. Major AraC/XylS-type regulators controlling the expression of degradation pathways for aromatic compounds
|
|
Most of the genes for XylS-type regulators lie upstream of their target operon, but, in contrast to lysR-type genes, they are transcribed in the same direction as the target genes. For example, cbdS, hpaA, cadR, benR, and phcT are all transcribed in the same direction as the genes they are regulating (36, 109, 196, 246). The pobR and pobC genes, on the other hand, are transcribed in the opposite direction (13, 173, 198). The xylS gene itself is also transcribed in the opposite direction to the meta-cleavage pathway operon (Fig. 1E) but is located downstream of it (93).
Expression of the genes for XylS-type regulators is not controlled just by themselves but in many cases by other activators or cascades. The best-studied example is XylS. Expression of xylS is strongly dependent on another regulatory protein XylR (see further below). XylR stimulates xylS-transcription from a 54-dependent promoter (called Ps1) when cells are grown on xylenes (69, 134). In the absence of suitable aromatic inducers, the xylS gene is expressed at low constitutive levels from a 70-dependent promoter called Ps2 (69). The hpaA regulatory gene has the unusual feature of being cotranscribed with the hpaX gene, which lies directly upstream of hpaA (194, 195). Expression of hpaA takes place from two promoters, one located directly upstream in the coding region of hpaX and the second located upstream of hpaX, in which case a bicistronic transcript is produced (195). Another example is provided by the phcT regulatory gene. PhcT is at the top of a regulatory cascade controlling phenol degradation (250). PhcT enhances transcriptional activation of the phc operon for phenol degradation by interacting with another regulator, PhcR, but at the same time it represses expression of the phcR gene itself. The dual action of PhcT is possible because its binding site is located in the intergenic region between phcR and the phc phenol operon, which themselves are oriented divergently (250).
Structure and Conformation
XylS/AraC-type regulators involved in degradation control of aromatic compounds are typically between 293 and 322 amino acids in size, with a molecular mass around 35 kDa. One exception is PhcT, which has only 257 amino acid residues with a predicted molecular mass of 28 kDa. As far as has been determined, most AraC members form dimers in solution (21, 24). The solution state of XylS has not been determined directly, since the protein is very difficult to purify. Ruiz et al. found that a peptide containing only the N-terminal end of XylS was able to dimerize (217). The C-terminal end of XylS/AraC-type regulators is the most highly conserved part among this protein family and contains two HTH motifs within a region of 100 or so amino acid residues (Fig. 2C) (71). In XylS the two HTH motifs comprise residues 231 to 252 and 282 to 305 (133), and a truncated XylS polypeptide with only the 112 C-terminal residues was capable of binding and activating transcription from the Pm promoter (103). Two AraC family members (i.e., Rob and MarA) have recently been crystallized (114, 210) (Fig. 6), but they do not match the most common AraC architecture, since MarA is a single-domain protein (with only a DNA binding domain) and Rob has the DNA binding domain located at the N terminus of the protein (52). MarA and Rob were crystallized as one monomer bound to a DNA fragment containing one binding site, CCAC-N7-TAA and GCAC-N7-CAA, respectively (Fig. 6). The two HTH motifs of MarA bind to adjacent segments of the major groove, with the helical axes of the recognition helices almost parallel to the DNA base pairs (Fig. 6). Because the recognition helices in the HTH motifs protrude from the same face of the protein, MarA binds to one face of the DNA. The two HTH motifs are connected by a rigid central linker helix that imposes their orientation and distance on the DNA binding site. Binding results in a bending of the DNA at an angle of 35°, because the two recognition helices are separated by 27Å whereas the pitch of the DNA B-form is 34 Å (210). The structure of the Rob DNA binding domain is quite similar to that of MarA, but two main differences were observed in the protein-DNA complex (Fig. 6). First, the DNA was not bent in the structure obtained with Rob, and second, only one protein helix was contacting the major groove of the DNA (114). In addition, nonspecifically attached Rob subunits were present in the crystallographic structure on the other side of the target DNA, which could be interpreted as an artifact (114). Martin and Rosner suggested that the presence of the second monomer prevented bending of the DNA and hence interaction of the second protein helix to the major groove of the DNA (137). On the other hand, it has been suggested that the Rob structure represents the state required for transcription activation, since conserved residues of one protein helix in the HTH-motif are not contacting the DNA but are available for interactions with the transcriptional machinery (114). Gallegos et al. had already previously suggested that conserved amino acids in this part of the HTH motif may contact the transcriptional machinery whereas the first helix would be more variable and involved in recognition of the target promoter (71).
View larger version (41K):
[in this window]
[in a new window]
|
FIG. 6. Structures of the Rob-micF and the MarA-mar DNA complex (114). The structurally similar N-terminal DNA binding domains of Rob and MarA are colored orange, and the unique C-terminal domain of Rob is shaded blue. (A) The N-terminal HTH motif of Rob contacting bases of the binding site. (B) MarA and the induced bend on its DNA binding site. Both HTH modules are situated in adjacent major-groove surfaces on one side of the DNA. Reprinted from reference 114 with permission from the publisher and from the authors.
|
|
The N-terminal end (also called the regulatory domain) of XylS/AraC members is not well conserved. The N-terminal domain of AraC has been crystallized in the absence and in the presence of arabinose, but it has such a low percent identity to XylS (12.5%) and XylS-mediated transcription activation from Pm is so different from AraC and the PBAD promoter (52, 71) that extrapolation of the AraC structure to XylS is doubtful. Structural ideas were retrieved, however, from genetic evidence and different XylS mutants. For example, single point mutations in the N-terminal end of XylS resulted in proteins responsive to other aromatic inducers, which suggests a role for the N-terminal end in effector recognition (Fig. 2C) (144, 218). Furthermore, some mutations in this region impaired the in vitro formation of cross-linking products between XylS peptides containing only the N-terminal part fused to the maltose binding protein. This was interpreted such that the N-terminal part of XylS would be involved in dimerization as well (217, 239). Both C- and N-terminal parts can act independently (218, 219). For example, mutations in both the C- and N-terminal regions of XylS can yield semiconstitutive mutants (203) or suppressors (143). This suggests there is some form of "cross talking" between the two domains, but the mechanism of this remains to be resolved.
Mechanisms of Activation
An understanding of the activation mechanisms by the "catabolic" members of the XylS/AraC family may be biased by the relatively large body of knowledge about XylS compared to others. A first peculiarity of the XylS system is that activation can take place in two modes, i.e., at low xylS expression levels in the presence of an inducer (66, 204) and on overexpression in the absence of inducer (91, 94, 142, 202, 241), although the latter mechanism may have no physiological meaning for the xyl pathway. At least BenR seems to behave similarly to XylS in this respect (36).
XylS probably binds as two monomers or one dimer to two direct sequence repeats in the Pm promoter. This can be concluded from crystallographic data for the two AraC members Rob and MarA (114, 210). DNase I footprinting of the XylS interaction on the Pm promoter showed a protected region from nucleotides –78 to –28 upstream of the transcription start site (102, 103). The protein bound along one side of the DNA, covering four helices and thereby making specific contacts in four adjacent major grooves (102). By earlier site-directed mutagenesis studies, a consensus direct repeat (TGCA-N6-GGNTA), which is present twice in the Pm promoter between –35 and –49 and between –56 and –70, had been identified as the binding determinant (70, 78, 107). In the presence of m-toluate, the degree of protection by XylS on the Pm promoter increased (102, 103), suggesting that the affinity of XylS to its binding site had changed. Subsequent in vivo methylation studies showed an altered methylation protection pattern of the XylS-bound region in the presence of m-toluate, which indicates that conformational changes take place under inducing conditions (78).
Strangely, XylS-mediated transcription activation from the Pm promoter is independent of 70-RNAP (136) but during exponential growth is dependent on the alternative sigma factor 32, which normally is involved in the heat shock response (135). This has led to the suggestion that the presence of the chemical effector m-toluate is triggering a similar response for the cell to the heat shock response. In the late exponential and stationary phases, the sigma requirement for the Pm promoter changes to 38 (135, 136). However, 38-RNAP uses exactly the same promoter region in the Pm promoter as that used by 32 (136). Mutants can be obtained with mutations in the N-terminal domain of XylS which change the requirement of XylS for 38 to 70 (203).
A large number of XylS/AraC members, including XylS itself, seem to require contact with -CTD and with parts of the -factor in order to achieve full activation (15, 53, 116, 117, 131, 220). This was concluded mostly from in vitro transcription experiments and transcriptional reporter gene fusions with mutated RNAPs. A direct interaction between XylS and RNAP has not been demonstrated conclusively (251), but introducing 2 bp between the nucleotides at positions –36 and –37 of the Pm promoter abolished XylS-dependent transcription activation, probably because XylS and RNAP were offset and could no longer interact (77). In vivo dimethyl sulfate footprinting suggested that XylS can retain RNAP on the Pm promoter in the absence of inducer. In the presence of benzoate, RNAP was released, with concomitant transcription initiation (147).
What can be concluded about the role of XylS in the process of transcription activation? Apparently, XylS prevents open-complex formation in the absence of inducer but somehow recruits RNAP to bind to the Pm promoter. In the presence of inducer, XylS changes its own conformation and that of its binding site, which allows isomeratization of promoter-RNAP to form the open complex (251). Exactly how the effector mediates these changes and how this is transmitted to the DNA and RNAP remain elusive.
Catabolic Operons Controlled by GntR-Type Regulators
Regulation by repression is rare for catabolic pathways. Ferrandez et al. were the first to report that the product of the paaX gene repressed the expression of two operons for phenylacetic acid (PA) degradation (paaABCDEFGHIJK and paaZ [Fig. 1E]) in E. coli strain W (63). These operons code for an aerobic catabolic pathway but have typical features of anaerobic degradation of aromatic compounds, because of activation of PA to phenylacetyl-coenzymeA (PA-CoA) by the action of a PA-CoA ligase (PaaK). By sequence similarity, the paaX gene product was classified as a GntR-type repressor. The paaX gene is located downstream of paaK and is transcribed in the same direction, although it is not part of the same transcript (63). Since then, a few other GntR members have been found for catabolic pathways (Table 4). The repressors PhcS and AphS regulate the expression of the phenol degradation pathway in C. testosteroni strains R5 and TA441 (7, 249). The phcS and aphS genes both belong to a small operon (phcRS and aphRS), of which the second gene is coding for an NtrC/XylR-type regulator (see below) (7). The aphRS and phcRS operons are located upstream of and transcribed in the opposite direction from their target catabolic operons (8, 249). The AphS protein acts as a factor for silencing the aph genes for phenol metabolism, whereas AphR is its transcription activator. Knockout mutations in the aphS gene were shown to enhance the transcription of the aph structural genes and of the aphR promoter as well (7). VanR, which represses the vanAB vanillate demethylase genes in Acinetobacter, is also a GntR-type regulator (149). The vanR gene is transcribed in the opposite direction to and overlaps with the vanB gene. Transcriptional regulation of biphenyl degradation in Ralstonia eutropha A5 (152), Pseudomonas sp. strain KKS102 (165), and P. pseudoalcaligenes KF707, is now also assumed to be mediated by GntR-type regulators (BphS) (269). The bphS gene in R. eutropha R5 is oriented in the opposite orientation to the first gene of the biphenyl operon (bphE) (152). In the strain KKS102, bphS is separated from its target operon by an insertion element (165). The third regulatory gene (orf0, now bphR2) present in a biphenyl catabolic regulon is also located upstream of the target operon but transcribed in the same direction (269). The Orf0 (now BphR2) protein is atypical in the GntR group since it is presumed to be an activator rather than a repressor.
GntR family members that control the degradation of aromatic compounds (except Orf0) are transcriptional repressors in the absence of the pathway substrates. In the presence of the pathway substrate, repression is released by interaction of the regulator with the aromatic compounds or one of its metabolites (62, 149, 165, 249). This was interpreted as a loss of DNA binding affinity mediated by the chemical inducer (see below) (62, 165). The mechanism by which Orf0 activates transcription is not known. Strong evidence for its activator function comes from gel shift assays performed with the orf0 promoter region and purified Orf0. In these experiments, OrfO bound the promoter fragment more strongly in the presence of inducers (2,3-dihydroxybiphenyl and 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoate) than in their absence (269).
Structure and Conformation
GntR-like regulators are between 239 and 254 amino acids long with a molecular mass of around 27 kDa. The PaaX protein is considerable bigger, with 316 amino acids and a molecular mass of 35 kDa. Structural information comes from GntR-type regulators that are not involved in pathways for degradation of aromatic compounds. Haydon and Guest first described bacterial regulators of the GntR family as having similar N-terminal DNA binding domains but displaying high heterogeneity among the various C-terminal effector binding and oligomerization domains (Fig. 2D) (83). This view has been largely confirmed by crystallographic structure determination and mutagenesis studies (199, 200, 259, 273). Exemplary for the GntR-type structure is FadR (211), whose structure has been resolved in the presence of its target DNA and its cognate inducer (Fig. 7). FadR exists as a homodimer in solution and binds DNA as a dimer (199, 259, 260, 273). The protein in the crystal structure showed an /ß N-terminal domain (residues 1 to 72) with the HTH motif and a C-terminal domain (residues 79 to 228) that is formed by seven -helices with short connecting loops. The seven -helices are packed together to form a bundle, and a large cavity within this bundle makes up the effector (myristoyl-CoA) binding pocket (259, 260). N- and C-terminal domains interact only via a short linker region comprising two short -helices. The two monomers in the dimer are packed together in parallel fashion, resulting in reciprocal interactions of the domains and linker regions across the interface (273). Only the tip of the paired -recognition helices projects orthogonally into the same major groove (259, 273). Each terminal portion of the -recognition helix interacts in an identical fashion with the DNA. Part of the HTH motif also penetrates the minor groove of the DNA helix to make specific contacts with two bases (259). The DNA has a B-form conformation with a curvature of 20° toward the protein, and this results in a contraction of the central major groove and an expansion of the opposite minor groove (273).
View larger version (67K):
[in this window]
[in a new window]
|
FIG. 7. Structure of the FadR-DNA complex (273). The paired recognition helices are in the major groove, and the two wings are within the flanking minor grooves. Note that the DNA is bent toward the FadR protein, leaving a slight contraction in the major groove as a result. Reprinted from reference 273 with permission from the publisher and from the authors.
|
|
Mechanisms of Regulation
The repression mechanism of GntR-type regulators seems to be a simple hindrance for RNAP binding or open-complex formation (Fig. 8). GntR-type proteins bind either the promoter region itself or between the transcription and translation start sites. For example, four regions on the bphE promoter were bound by purified BphS of Pseudomonas sp. strain KKS102. The binding sites are located between positions –12 and +3 (called BSI), +6 and +23 (BSII), +27 and +40 (BSIII), and +44 and +58 (BSIV) compared to the transcription start site. BphS has more affinity to BSI and BSII than to BSIII and BSIV, a difference which was explained by the presence of inverted repeats in BSI and BSII (Table 4), and no such features in BSIII and BSIV. In fact, fragments with only BSI and BSII are sufficient for BphS to mediate repression, whereas transcription from an artificial promoter fragment containing only BSIII and BSIV was not repressed to a significant extent. Nevertheless, the strongest repression was obtained when all four binding sites were present (165). BphS loses its ability to bind to the promoter in the presence of the inducer 2-hydroxy-6-oxo-6-phenylhexa-2,4-dienoic acid (an intermediate in biphenyl metabolism).
View larger version (22K):
[in this window]
[in a new window]
|
FIG. 8. Two modes of repression by GntR-type regulators. (A) The repressor bound to the –10 promoter region impairs RNA polymerase binding. (B) Binding of an inducer abolishes DNA binding upon which RNA polymerase can start transcription. (C) Repressor bound between the transcription and translation start site impairs open-complex formation. (D) Binding of the inducer to the repressor abolishes DNA binding activity. Regulatory proteins are shown as ellipsoids.
|
|
A 15-bp imperfect palindromic sequence was identified in the PaaX-dependent promoters paaZ and paaA. The palindrome was located between positions –30 and –16 in the paaZ promoter and between +3 and +17 in the paaA promoter. DNase I footprinting showed that the PaaX protein actually protects a larger region, which extends from –1 to +51 in the paaA promoter and from –30 to +17 in the paaZ promoter (62). Hypersensitive sites were spaced at approximately 10-nucleotide intervals, suggesting binding of PaaX to one side of the double helix. The fact that PaaX protected the –10 box in the paaZ promoter but not in the paaA promoter suggests a different mechanism of PaaX repression in these two promoters (as shown in the model in Fig. 8). In the presence of phenylacetate-CoA, binding of PaaX to the paaA and paaZ promoters was released (62). A sequence of events leading to the release of repression can be deduced from the FadR-DNA structure in the presence of myristoyl-CoA (the inducer for FadR) (259). Binding of the effector induces conformational changes in the sensor kink helix 8 (Fig. 7), which then forms two separate helices. As a result, some of the side chains on this helix are pushed toward the neighboring helix, 4 (linker region), which responds by shifting away from helix 8 toward the N-terminal DNA binding domain. This shift introduces additional contacts with the helix 1 in the N-terminal domain, which tilts away and causes a hinge-bending motion of both DNA binding domains of the dimer in opposite direction. This leads to a change in the distance between the two recognitions helices, which now lose their ability to interact with the DNA helix (259).
|
TetR-, MarR-, AND FNR-TYPE REGULATORS
|
|---|
TetR Family
Only one example of a TetR-type regulator involved in the regulation of an catabolic pathway for aromatic compounds is known. This protein, CymR, is a repressor for the expression of the genes for p-cymene (cym) and p-cumate (cmt) degradation in P. putida F1 (Fig. 1H; Table 5) (50). The cymR gene itself is positioned upstream of and transcribed in the same direction as the cym operon. Similar sequences have been identified in the cmt and cym promoter regions (Table 5) and may be the recognition sites for the CymR repressor protein (50, 164). CymR is supposed to impose its repressing effect by inhibiting RNAP access to the promoter. In the presence of the pathway effector, p-cumate, CymR is assumed to no longer bind to its operator site, in analogy to the TetR repressor in the presence of tetracycline (170). Strangely, p-cymene, the substrate for the cym-encoded pathway, is not an effector for CymR (50). The CymR protein is 203 amino acids long with a molecular mass of 23 kDa and is supposed to form a dimer in solution, like the TetR protein, for which the structure is known (170). Its N-terminal part contains an HTH motif (predicted for amino acids 19 to 64 in CymR), which is responsible for DNA binding, whereas the C-terminal domain binds tetracycline (170). The C-terminal domains interact in the dimer to form a core. The HTH motifs bind to the (palindromic) operator sequence, with the recognition helices aligned parallel to the major groove. The distance and the relative orientation of the HTH motifs distinguish between the induced and the noninduced status of TetR. Inducer binding triggers a sequence of -helix displacements, which results in an increasing separation of the recognition helices. The shift of the recognition helices along the major groove disrupts the contacts between the DNA binding domains and the operator, causing dissociation of the TetR operator complex (170).
View this table:
[in this window]
[in a new window]
|
TABLE 5. TetR, MarR, and FNR family members and two-component regulators in degradation pathways of aromatic compounds
|
|
MarR-Type Regulators
The nbz operon for aminophenol degradation on plasmid pNB1 in P. putida HS12 is regulated by NbzR, a MarR-type repressor protein (176). The chemical inducer for the pathway has not been identified, and it was proposed that another trans-acting factor might be involved in regulating pathway expression. Other MarR-type regulators involved in catabolic pathways include HpcR (HpaR), the repressor of the hpc (hpa) operon for 3- and 4-hydroxy- and 3,4-dihydroxyphenylacetic acid (homoprotocatechuate) degradation in E. coli strains C (214) and W (68). In fact, current information from genome sequences shows that the pathways for homoprotocatechuate degradation are very common among Gamma- and Betaproteobacteria. CbaR, another MarR-type regulator, controls the cbaABC operon for 3-chlorobenzoate degradation present on plasmid pBRC60 in C. testosteroni BR60 (197). 3-Chlorobenzoate and protocatechuate are effectors for CbaR, leading to derepression (197). The genes for benzoyl-CoA reductase in Rhodopseudomonas palustris are also under the control of a MarR family member, named BadR. BadR activates rather than represses gene expression (55), although most other MarR family members are repressors of gene transcription (113, 171). Benzoate is assumed to be the effector for BadR-mediated activation (55). Additional pathway control is exerted by the AadR regulator, a FNR-type regulator, under anaerobic conditions (55). The regulatory genes nbzR, cbaR, and hpcR are located upstream of and are transcribed in the opposite direction to their target operons (176, 194, 197). badR is located upstream of and transcribed in the same direction as its target operons but is separated from them by one additional gene (55). HcaR, the repressor of the hydroxycinnimate (hca) genes in Acinetobacter sp. strain ADP1, is also a distant member to MarR (28% identical aligned residues), which reacts to a hydroxycinnamoyl-CoA thioester as an effector compound (180).
DNA binding studies have been performed with HpaR (68) and CbaR (197). CbaR contacts two sites (named BSI and BSII), which are located near the transcription start site. BSI is located approximately 40 nucleotides downstream of the transcription start site of cbaA, whereas BSII overlaps with it. An inverted repeat of 4 nucleotides separated by 6 and 9 nucleotides (GTTG-N6/9-CAAC) is present in the BSI region. Similar (but not perfect) inverted repeats are present in BSII (GTTG-N6-TAAC or GTAG-N9-TAAC), and it is presumed that CbaR interacts with the inverted-repeat sequences in some way. In DNase I footprinting experiments, CbaR bound to a BSI DNA fragment with higher affinity than to BSII, which led to the conclusion that independent CbaR proteins bind to each site. In the presence of 3-chlorobenzoate, benzoate, or protocatechuic acid, DNA binding by CbaR is disrupted, whereas 3-hydroxy- and 3-carboxybenzoate increased binding affinity by 60-fold. The physiological significance of this phenomenon is not understood (197). HpaR from E. coli is binding to two operator sites in the Pg promoter for the hpaGEDEFHI operon, one of which overlaps with the transcriptional start site (68). The operator around the transcriptional start site contains a perfect palindromic sequence (AATCATTAA-N4-TTAATGATT). Transcription from the Pg promoter is furthermore subject to strong catabolite repression by the cyclic AMP receptor protein.
Most MarR-type proteins carry a conserved HTH motif in the center of the protein, whose function would be to bind DNA (4). NbzR is atypical, since several conserved residues of the HTH motif are not present in NbzR. Instead, the protein possesses a putative leucine zipper motif (which is absent in other family members), which is thought to be responsible for the NbzR dimerization and DNA binding. Hence, Park and Kim suggested that the poor HTH motif of NbzR perhaps does not contribute to DNA binding (176). MarR-type regulators controlling catabolic operons are relatively small proteins. They contain between 148 and 196 amino acids and have a molecular mass of between 17 and 22 kDa. MarR has been crystallized (4, 130, 272). The structure revealed a MarR dimer with each subunit consisting of six helical regions. The N- and C-terminal regions encompassing residues 10 to 21 and 123 to 144, respectively, are closely juxtaposed and intertwine with the equivalent regions of the second subunit to form a domain that holds the dimer together (Fig. 9). These N- and C-terminal domains are linked to the remainder of the protein by two long antiparallel helices in each subunit. The helices lead to one globular domain, which includes residues 55 to 100 of each subunit and might be responsible for DNA binding. Although the globular DNA binding domains of the dimer are adjacent, they make minimal contact with each other. Four molecules of salicylate, which is an effector for MarR, were visible in the crystal structure: two at each subunit and two on either side of the proposed DNA binding helix (4).
FNR-Type Regulators
Three members of the FNR family of regulators were found to be involved in controlling catabolic operon expression (Table 5). These include HbaR, which activates the expression of the gene for 4-hydroxybenzoate-CoA ligase in response to 4-hydroxybenzoate in R. palustris (56); AadR, also present in R. palustris (55); and CprK of Desulfitobacterium dehalogenans (238). AadR is the transcriptional activator for the hbaR gene itself. It is also, with BadR (see above), a coactivator for the badD benzoyl-CoA reductase gene in R. palustris (55, 56), functioning as an oxygen sensor and enhancing the expression of the anaerobic degradation pathway under oxygen exclusion. Upstream of the transcription start sites of hbaR and badD, interrupted inverted repeats (TTGAT-N4-ATCAA) are found, which were postulated to be AadR binding sites (55, 56). The cprK gene is part of a cluster for o-chlorophenol dehalogenase (238). One small transcript which encompasses the cprBA genes for the dehalogenase itself is induced under conditions of dehalorespiration; the remainder of the cpr genes are expressed constitutively. At several points within the cpr cluster, FNR-like binding motifs were detected. Within the <
Top
Next
