INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Within the last few decades there has been a notable increase in government regulations that hold industries accountable for the chemical pollution that results from their manufacturing activities. In order to comply with environmental laws, industries must be able to identify contamination resulting from chemical spills and leaks and to monitor remediation processes. The chromatographic methods currently used for analysis of samples from contaminated sites are costly and technically complex (14). A potentially inexpensive and simple way to reduce the cost of contaminant detection is to use biosensors derived from the genetic systems of bacteria that use organic contaminants as growth substrates.

A whole-cell bacterial biosensor can be created by placing a reporter gene under control of an inducible promoter. Expression of the reporter gene provides a measurable response when the appropriate transcription activator protein interacts with an effector molecule to signal a particular environmental condition. Bacterial biosensors have been created to detect naphthalene (10), benzene derivatives, including toluene and xylene (1, 11, 20, 41); and certain toxic metals (20, 25).

The construction and use of bacterial biosensors has been restricted by our limited understanding of the genetic systems that control bacterial responses to polluting chemicals and by the specificity of the interaction between a transcription activator protein and its chemical effectors. However, of the known systems, several show a high degree of similarity in the regulatory pathways that control their expression. Operons carrying genes required for metabolism of phenol, toluene, benzene, and xylene in some Pseudomonas and Acinetobacter species are controlled through inducible promoters recognized by ⁵⁴-associated RNA polymerase. Transcription directed by these promoters occurs when a regulatory protein detects the presence of the substrate for the catabolic enzymes. Proteins in this category are DmpR, XylR, MopR, PhhR, PhlR, and TbuT (3, 6, 16, 17, 24, 28, 29, 31). These six proteins show significant similarity to one another at the amino acid level and are organized into discrete domains with independent functions that include chemical detection, polymerase activation, and DNA binding.

XylR and DmpR are the best-characterized members of this group of transcription activators. The Pseudomonas putida XylR protein has served as the detection component for a number of biosensors (1, 11, 20, 41) based on the ability of this protein to activate transcription in response to xylene and toluene. DmpR, the product of the Pseudomonas sp. strain CF600 dmpR gene (28, 29), mediates expression of the dmp operon to allow growth on simple phenols. Transcription from Pdmp, the promoter of the dmp operon, is activated when DmpR detects the presence of an inducing phenol (29).

Domain-swapping experiments to form XylR-DmpR hybrids demonstrated that the sensor activity of these regulatory proteins is localized in the amino-terminal region (29). Transcription from Pdmp depends on a direct physical interaction between the sensor domain of DmpR and the inducing phenol. A productive association between the sensor domain and a phenolic molecule causes DmpR to undergo a conformational change that results in a polymerase-activating form of the protein (18, 30).

The single-protein, independent-domain arrangement of DmpR and other proteins of this type makes them particularly suitable candidates for genetic manipulation. Specifically, one should be able to alter the chemical-sensing domain of the protein through mutagenesis without disturbing DNA-binding or transcription-activating functions. Modification of the sensor domain should allow the creation of novel proteins that respond to xenobiotics which remain undetected by the wild-type protein. Such altered proteins have the potential to extend the chemical target range of biosensors beyond that based on natural systems.

The natural interaction of DmpR with a subset of phenols suggested that modification of its sensor domain might create protein derivatives with the capacity to detect phenolic molecules commonly used in industry. Phenol and substituted phenols are common starting materials and waste by-products in the manufacture of chemical, industrial, and agricultural products (34-39). The high-volume use of phenols in the United States and their potential toxicity has led the U.S. Environmental Protection Agency to include 11 of them on its list of priority pollutants. Phenols listed as priority pollutants include phenol, 2-chlorophenol, 2,4-dichlorophenol, 2,4,6-trichlorophenol, pentachlorophenol, 4-chloro-3-methylphenol, 2,4-dimethylphenol, 2-nitrophenol, 4-nitrophenol, 2,4-dinitrophenol, and 2-methyl-4,6-dinitrophenol (9).

The magnitude of the production and use of phenolic molecules in the United States suggests that an inexpensive method of detecting these chemicals could be of use in preventing the inadvertent release of phenolic pollutants. Here, we present five DmpR mutants that were created through random mutagenesis of the DmpR sensor domain. All five of the DmpR mutants described in this report, through activation of a reporter gene, increase detection of the phenolic effectors that also generate transcription activation by wild-type DmpR. In addition, the DmpR mutants demonstrate the ability to detect phenolic molecules that are not effectors of the wild-type protein. The distribution of mutations that alter the chemical-sensing capacity of DmpR suggests that protein secondary structure, determined by the amino acid sequence of the sensor domain, normally prevents disubstituted phenols from acting as effectors of DmpR-mediated transcription.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial strains and plasmids. The bacterial strains and plasmids used in this study are listed in Table 1. Escherichia coli TE2680 (8) was used as an intermediate strain for placing the Pdmp-lacZ fusion into the chromosome of E. coli MC4100 (5) to create a test strain (AW101) for the DmpR derivatives. DH5 (27) was the host for plasmid constructions.

Genetic techniques. Plasmid DNA was isolated by using a Plasmid Kit (Qiagen, Inc., Chatsworth, Calif.) or by a miniprep alkaline lysis method (13). Standard methods were used for restriction digests, gel electrophoresis, and ligations (2, 27). Transformation of E. coli was accomplished by electroporation (7) with Gene Pulser II unit (Bio-Rad, Hercules, Calif.). Nonmutagenic PCR to amplify the Pdmp fragment was done as described by Innes et al. (12).

Construction of DmpR derivatives with mutant sensor domains. The method used to create and select mutant DmpR proteins with increased responses to phenol and substituted phenols is diagrammed in Fig. 1. Mutagenic PCR of the DmpR sensor domain was conducted by a modification of Cadwell and Joyce's method (4). Plasmid pAW50 served as the template in the mutagenic PCR with 25 pmol each of primers dmpR5'-75 (5'-GCCGTCGATTGATCATTTGG-3') and dmpR3'-976 (5'-TGTCCATCATATTGCGCACG-3'). In addition, the reaction mixture contained 5 mM MgCl₂, 0.5 mM MnCl₂, 0.2 mM dATP and dGTP, 0.8 mM dCTP and dTTP, 10 mM Tris-HCl (pH 8.3), 50 mM KCl, 0.001% (wt/vol) gelatin, and 5 U of AmpliTaq polymerase (Perkin-Elmer). The mutagenic PCR amplification cycle followed a 2-min denaturation at 92°C and consisted of 30 cycles of 94°C (10 s), 56°C (20 s), and 72°C (1 min).

View larger version (16K): [in this window] [in a new window]   FIG. 1.   Independent domain structure of dmpRwt-N and method used to create and select DmpR mutants. aa, amino acid.

Test strain construction and screen for sensor domain mutations. The pAW51 plasmid, carrying the Pdmp-lacZ fusion, was linearized by restriction with ScaI and then used to transform TE2680 with selection for kanamycin resistance. Transformants were screened for loss of ampicillin and chloramphenicol resistance, a condition indicating integration of the Pdmp-lacZ fusion into the TE2680 chromosome at the trp operon (8, 32). The general transducing phage P1kc (American Type Culture Collection, Manassas, Va.) was used to transfer the fusion to the chromosome of E. coli MC4100. The transduction created strain AW101.

DmpR mutant selection. AW101 was used as a test strain to identify and characterize changes in the chemical detection capacity of DmpR derivatives after sensor domain mutagenesis. The pAW50 plasmid derivatives were introduced into AW101 by electroporation. Transformants were selected on Luria-Bertani medium (Difco, Detroit, Mich.) plates containing 10.5 µg of tetracycline per ml. Transformants were then replica plated onto M9 minimal medium (15) plates containing 0.2% glucose, 30 µg of tryptophan per ml, 1 µg of thiamine per ml, 10.5 µg of tetracycline per ml, 0.003% 5-bromo-4-chloro-3-indoyl--D-galactoside (X-Gal), and either no inducer or a phenol derivative at 50 nM. Cells that formed bluer colonies than cells containing wild-type DmpR were subjected to further analysis with liquid -galactosidase assays.

Phenolic molecules tested. DmpR mutants were tested for their ability to detect the 11 phenols listed as priority pollutants by the U.S. Environmental Protection Agency, i.e., phenol, 2-chlorophenol, 2,4-dichlorophenol, 2,4,6-trichlorophenol, pentachlorophenol, 4-chloro-3-methylphenol, 2,4-dimethylphenol, 2-nitrophenol, 4-nitrophenol, 2,4-dinitrophenol, and 2-methyl-4,6-dinitrophenol. Phenols were prepared as 25 mM stocks in ethyl alcohol for addition to solid or liquid media and were tested at concentrations ranging from 0.1 to 75 µM in different experiments.

-Galactosidase assays. Overnight cultures of AW101 carrying pAW50 derivatives were diluted 1,000-fold into Luria-Bertani broth containing 10.5 µg of tetracycline per ml. When cells reached an A₅₉₅ of between 0.80 and 1.0 as measured on a Lambda Bio UV-visible spectrophotometer (Perkin-Elmer Corp. Analytical Instruments, Norwalk, Conn.), 800-µl samples were pelleted by centrifugation and immediately suspended in 800 µl of spent Luria-Bertani broth (medium from an overnight culture sterilized through a 0.2-µm-pore-size filter) containing the appropriate phenol derivative. The use of spent Luria-Bertani medium prevented the high level of phenol-independent activity that was seen when cells were allowed to grow into stationary phase in regular Luria-Bertani medium or glucose-supplemented minimal medium. In E. coli, the dmpR gene is expressed or stabilized as cells enter stationary phase (33). We assume that the stationary-phase expression of the Pdmp-lacZ fusion in the absence of a phenolic inducer resulted from the multicopy nature of pAW50 and its derivatives carrying the dmpR mutants.

DNA sequencing. Mutations in the dmpR sensor domain were identified through DNA sequencing with the ABI PRISM Dye Terminator Cycle Sequencing system (Perkin-Elmer Corp.). Products from sequencing reactions were separated by electrophoresis in 4% polyacrylamide gels on an ABI 373A Stretch DNA Sequencer (Applied Biosystems, Inc., Foster City, Calif.). Analysis of mutant sensor domain DNA and amino acid sequences was conducted by using LASERGENE software (DNASTAR Inc., Madison, Wis.).

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

DmpR mutants increase detection of phenolic molecules in -galactosidase assays. Our mutagenic procedure and selection process (Fig. 1) resulted in replacement of DNA corresponding to the first 176 amino acids of DmpR (about 85% of the sensor domain). Three repetitions of these methods led to the identification of more than 20 DmpR derivatives that increased transcription of the Pdmp-lacZ fusion (compared to wild-type DmpR) in response to phenol and substituted phenols on plate assays. Sequence analysis demonstrated that the number of mutations per dmpR derivative ranged from one to six and averaged between two and three (data not shown).

Mutations that change the interaction of DmpR with phenol and substituted phenols. Changes in the sensor domain-coding sequences of the DmpR mutants that increased transcription of the Pdmp-lacZ fusion in response to phenol or substituted phenols are represented in Fig. 2. Three DmpR mutants contained a mutation at codon 116 and/or codon 117. B23 contained a lysine-to-methionine change at codon 117, F17 contained an aspartate-to-valine change at codon 116, and D9 was mutated at both positions (D116G and K117R). B23 and D9 contained additional mutations, making it difficult to assess the importance of changes at codon 117. An enhanced response to phenolic molecules mediated by F17 (D116V) suggested that the aspartate at position 116 acts to restrict the effector range of wild-type DmpR. The Q10R mutation of B24 strongly enhanced the response to phenol and disubstituted phenols. The effector response profile of B24 is similar to that of mutant B23, suggesting that the Q10R mutation contributes to increased detection of phenols by B23.

View larger version (16K): [in this window] [in a new window]   FIG. 2.   Amino acid sequences of DmpR mutant sensor domains. The wild-type sequence is shown in the first line. Mutations are entered at corresponding positions for the mutants. The # symbol indicates a nucleotide mutation that is not expected to alter the amino acid code (e.g., TTTTTC at codon 48 in D9).

Detection of phenolic molecules by DmpR mutants. The DmpR mutants described in this report were characterized in detail with liquid -galactosidase assays. -Galactosidase activity is proportional to transcription of the Pdmp-lacZ reporter fusion and is, therefore, a measure of a particular DmpR mutant's ability to detect phenol or a specific substituted phenol. The comparative performance of mutant proteins was initially tested to detect priority pollutant phenols listed by the U.S. Environmental Protection Agency in concentrations that ranged from 10 to 100 µM. The concentrations of phenolic effectors for the assays represented in Tables 2 and 3 were chosen based on their ability to elicit a response from the majority of DmpR mutants that was at least fourfold higher than that achieved in the absence of a phenolic inducer. Thus, phenol, 2-chlorophenol, and 2,4-dichlorophenol were used at a concentration of 25 µM (Table 2). The less efficient inducers 4-chloro-3-methylphenol, 2,4-dimethylphenol, 2-nitrophenol, and 4-nitrophenol (Table 3) were used at 75 µM.

Priority pollutant phenols that are not detected by DmpR mutants. Priority pollutant phenols that were not detected by any of the DmpR mutants included the larger and more toxic molecules, i.e., pentachlorophenol, 2,4,6-trichlorophenol, and 2-methyl-4,6-dinitrophenol. The pK_as of pentachlorophenol (4.9) and 2-methyl-4,6-dinitrophenol (4.3) suggest that these molecules may not have been able to efficiently enter the cell due to deprotonation of their hydroxyl groups at the neutral pH of the induction medium. In addition, exposure of the test strain to 25 µM pentachlorophenol or 2,4,6-trichlorophenol resulted in visible cell lysis. The response to 2,4-dinitrophenol was unaccountably variable.

Limits of phenol detection by DmpR mutant B24. The usefulness of a DmpR variant as the phenol-detecting component of a bacterial biosensor will be determined in part by the lowest concentration of chemical that can elicit a transcriptional response. As illustrated in Tables 2 and 3, the B24 mutant protein was one of the more sensitive DmpR derivatives. We sought to estimate the lowest concentrations of phenol, 2-chlorophenol, and 2,4-dichlorophenol that could be detected by B24. Reliable detection was defined as -galactosidase activity that reached a level at least fourfold higher than that attained in the absence of a phenolic inducer. In order to amplify the -galactosidase signal, the chemical exposure time was increased to 4 h.

DmpR derivatives do not respond to XylR effectors. The sensor domain of XylR is 64% identical to DmpR at the amino acid level (28). The increased range of phenol effectors for the DmpR derivatives led us to ask if they had acquired the ability to respond to effectors of XylR. Cells carrying either wild-type DmpR or mutant B9, B23, D9, or F17 were exposed to 75 mM toluene, 2-nitrotoluene, 2-chlorotoluene, 4-chlorotoluene, or o-xylene in spent Luria-Bertani medium for 2 h. These chemicals did not induce transcription of the Pdmp-lacZ fusion (data not shown). Thus, like the wild-type protein, the DmpR mutants were unable to respond to chemical effectors of XylR.

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

Whole-cell bacterial biosensors have the potential to provide inexpensive, easy-to-use methods of detecting industrial pollution. Biosensors based on the genetic systems of bacteria that metabolize pollutants have been described (1, 10, 11, 20, 40, 41). These biosensors have used the XylR or NahR proteins to activate expression of reporter genes. The chemical detection capabilities of bacterially based biosensors have been limited to chemicals that are natural effectors of a wild-type regulator protein. Through the generation and characterization of mutant regulator proteins, we demonstrate that the chemical detection ability of bacterial biosensors can be enhanced to increase the sensitivity to chemical effectors of the natural genetic system. The range of effector chemicals that can be detected by the bacterial system can also be expanded to include chemicals that are not detected by the wild-type proteins.

In this report we describe the response of DmpR derivatives to phenols listed as priority pollutants by the U.S. Environmental Protection Agency. DmpR mutants B9, B23, B24, and F17 increased sensitivity to effectors of the wild-type protein and also responded to phenolic molecules that are not effectors of the natural system. Engineered proteins of this type could form the basis for biosensors with improved detection of organic pollutants.

The mechanism by which DmpR binds its chemical effectors and changes conformation to become capable of transcription activation is not well understood. However, there is good evidence that the capacity of DmpR to activate transcription is repressed through an interaction between the sensor and polymerase-activating domains of DmpR (18, 19, 30). This repression can be reversed through a successful interaction between the sensor domain and an effector chemical or through engineering the removal of the sensor domain (30). Similar experiments with XylR give analogous results (22, 23).

Our DmpR mutants failed to respond to XylR effectors (data not shown). The benzene derivatives that induce transcription activation mediated by XylR are similar in size and shape to the phenolic effectors studied, but they lack a hydroxyl group. Our results suggest that a hydroxyl group remains a requirement for achieving the transcription activator form of the DmpR mutants. It is clear, however, that DmpR mutants B9, B23, B24, D9, and F17 can accommodate an increased number of substituents on their phenolic effectors compared with the wild-type protein.

Sensor domain mutations that increase the range of effector chemicals have been identified by others in an attempt to understand the biochemical properties of DmpR and XylR. DmpR mutations E135K and R184W each induce transcription of a reporter gene following exposure to phenolic molecules that are not effectors of wild-type DmpR (21, 30). Like the DmpR derivatives described in this paper, these mutant proteins retain the capacity to respond to chemical effectors of wild-type DmpR. An XylR E172K mutation allows a response to 3-nitrotoluene, a chemical which inhibits transcription activation by the wild-type protein (6, 26). It has been suggested that mutations that alter the effector profile of DmpR or XylR exert their effect either through an improvement in the effector-protein interaction or by changing the three-dimensional structure of the protein in ways that enhance other necessary functions of the protein, for example, polymerase activation (6, 21, 26).

The fact that the DmpR derivatives can still detect effectors of wild-type DmpR (phenol, 2-chlorophenol, and 2-nitrophenol) suggests that a key aspect of the interaction between the mutant sensor domains and the chemical effectors remains substantially unaltered. Rather, mutations that distort the secondary structure of DmpR may allow larger, differently substituted molecules access to a crucial site on the sensor domain. These molecules could then act as effectors, provided that they have the appropriate chemical attribute (perhaps determined by the phenol hydroxyl group).

For DmpR and similarly organized proteins, knowledge of the protein's three-dimensional structure and any specific contacts it makes with effector chemicals would greatly enhance efforts to directly engineer proteins that can function as detectors for a variety of organic molecules. So far, the poor solubility of DmpR and related proteins has impeded efforts to understand the structure of such proteins.