Search

The Vibrio fischeri LuxR protein is the founding member of a family of acyl-homoserine lactone-responsive quorum-sensing transcription factors. Previous genetic evidence indicates that in the presence of its quorum-sensing signal, N-(3-oxohexanoyl) homoserine lactone (3OC6-HSL), LuxR binds to lux box DNA within the promoter region of the luxI gene and activates transcription of the luxICDABEG luminescence operon. We have purified LuxR from recombinant Escherichia coli. Purified LuxR binds specifically and with high affinity to DNA containing a lux box. This binding requires addition of 3OC6-HSL to the assay reactions, presumably forming a LuxR-3OC6-HSL complex. When bound to the lux box at the luxI promoter in vitro, LuxR-3OC6-HSL enables E. coli RNA polymerase to initiate transcription from the luxI promoter. Unlike the well-characterized LuxR homolog TraR in complex with its signal (3-oxo-octanoyl-HSL), the LuxR-30C6-HSL complex can be reversibly inactivated by dilution, suggesting that 3OC6-HSL in the complex is not tightly bound and is in equilibrium with the bulk solvent. Thus, although LuxR and TraR both bind 3-oxoacyl-HSLs, the binding is qualitatively different. The differences have implications for the ways in which these proteins respond to decreases in signal concentrations or rapid drops in population density.

LuxR is the transcriptional activator for quorum-sensing control of luminescence in Vibrio fischeri. A series of alanine-scanning mutants spanning a predicted helix-turn-helix region in the DNA binding domain of LuxR was constructed, and the activity of each of the LuxR mutant proteins in recombinant Escherichia coli was investigated. The region covered by the mutagenesis spanned residues 190 to 224. About half of the alanine-scanning mutants showed activities similar to that of the wild-type LuxR: at least two were positive-control mutants, four appeared to be defective in DNA binding, and several others were characterized as DNA binding affinity mutants. This analysis, taken together with information about other bacterial transcription factors, provides insights into amino acid residues in LuxR that are involved in DNA binding and transcriptional activation.

The Vibrio fischeri luminescence (lux) operon is regulated by a quorum-sensing system that involves the transcriptional activator (LuxR) and an acyl-homoserine lactone signal. Transcriptional activation requires the presence of a 20-base inverted repeat termed the lux box at a position centered 42.5 bases upstream of the transcriptional start of the lux operon. LuxR has proven difficult to study in vitro. A truncated form of LuxR has been purified, and together with sigma(70) RNA polymerase it can activate transcription of the lux operon. Both the truncated LuxR and RNA polymerase are required for binding to lux regulatory DNA in vitro. We have constructed an artificial lacZ promoter with the lux box positioned between and partially overlapping the consensus -35 and -10 hexamers of an RNA polymerase binding site. LuxR functioned as an acyl-homoserine lactone-dependent repressor at this promoter in recombinant Escherichia coli. Furthermore, multiple lux boxes on an independent replicon reduced the repressor activity of LuxR. Thus, it appears that LuxR can bind to lux boxes independently of RNA polymerase binding to the promoter region. A variety of LuxR mutant proteins were studied, and with one exception there was a correlation between function as a repressor of the artificial promoter and activation of a native lux operon. The exception was the truncated protein that had been purified and studied in vitro. This protein functioned as an activator but not as a repressor in E. coli. The data indicate that the mutual dependence of purified, truncated LuxR and RNA polymerase on each other for binding to the lux promoter is a feature specific to the truncated LuxR and that full-length LuxR by itself can bind to lux box-containing DNA.

Acylhomoserine lactones, which serve as quorum-sensing signals in gram-negative bacteria, are produced by members of the LuxI family of synthases. LuxI is a Vibrio fischeri enzyme that catalyzes the synthesis of N-(3-oxohexanoyl)-L-homoserine lactone from an acyl-acyl carrier protein and S-adenosylmethionine. Another V. fischeri gene, ainS, directs the synthesis of N-octanoylhomoserine lactone. The AinS protein shows no significant sequence similarity with LuxI family members, but it does show sequence similarity with the Vibrio harveyi LuxM protein. The luxM gene is required for the synthesis of N-(3-hydroxybutyryl)-L-homoserine lactone. To gain insights about whether AinS and LuxM represent a second family of acylhomoserine lactone synthases, we have purified AinS as a maltose-binding protein (MBP) fusion protein. The purified MBP-AinS fusion protein catalyzed the synthesis of N-octanoylhomoserine lactone from S-adenosylmethionine and either octanoyl-acyl carrier protein or, to a lesser extent, octanoyl coenzyme A. With the exception that octanoyl coenzyme A served as an acyl substrate for the MBP-AinS fusion protein, the substrates for and reaction kinetics of the MBP-AinS fusion protein were similar to those of the several LuxI family members previously studied. We conclude that AinS is an acylhomoserine lactone synthase and that it represents a second family of such enzymes.

LuxR is a sigma(70) RNA polymerase (RNAP)-dependent transcriptional activator that controls expression of the Vibrio fischeri lux operon in response to an acylhomoserine lactone-cell density signal. We have investigated whether the alpha-subunit C-terminal domain (alphaCTD) of RNAP is required for LuxR activity. A purified signal-independent, LuxR C-terminal domain-containing polypeptide (LuxRDeltaN) was used to study the activation of transcription from the luxI promoter in vitro. Initiation of lux operon transcription was observed in the presence of LuxRDeltaN and wild-type RNAP but not in the presence of LuxRDeltaN and RNAPs with truncated alphaCTDs. We also studied the in vivo role of the RNAP alphaCTD in activation of lux transcription in Escherichia coli. This enabled a comparison of results obtained with full-length LuxR to those obtained with LuxRDeltaN. These in vivo studies indicated that both LuxR and LuxRDeltaN require the RNAP alphaCTD for activity. The results of DNase I protection studies showed that LuxRDeltaN-RNAP complexes can bind and protect the luxI promoter, but with less efficacy when the alphaCTD is truncated in comparison to the wild type. Thus, both in vitro and in vivo experiments demonstrated that LuxR-dependent transcriptional activation of the lux operon involves the RNAP alphaCTD and suggest that alphaCTD-LuxR interactions may play a role in recruitment of RNAP to the luxI promoter.

Although cell density-dependent regulation of the luminescence genes in Vibrio fischeri is a model for quorum sensing in Gram-negative bacteria, relatively little is known about the promoter of the luminescence operon. The luminescence operon is activated by the LuxR protein, which requires a diffusible acylhomoserine lactone signal. The lux box, a 20 bp inverted repeat, is located in the luxl promoter region and is required for LuxR-dependent induction of the luminescence genes. Using primer extension, we mapped the LuxR-dependent transcriptional start site of the lux operon to 19 bp upstream of the luxl start codon. This indicates that the lux box is centred at -42.5 bp from the start of transcription. To gain evidence about the location of the -10 sequence, we placed a consensus -35 hexamer at different locations relative to the luxl transcriptional start site and measured constitutive levels of luminescence in recombinant Escherichia coli. The strongest constitutive promoter contained a TATAGT hexamer 17 bp from the -35 consensus sequence and 6 bp from the transcriptional start site. We propose that this is the -10 hexamer. Also in recombinant E. coli, both half-sites of the lux box were required for LuxR-dependent gene activation and for activation by an autoinducer-independent, monomeric LuxR deletion protein. LuxR-dependent activation of luminescence was eliminated when the lux box was centred at -47.5, -52.5 and -62.5 with respect to the luxl transcriptional start site. Our evidence, taken together with other information, points to a model in which a LuxR dimer overlaps the -35 region of the luxl promoter and functions as an ambidextrous activator with each LuxR subunit interacting with a different region of RNA polymerase.

Synthesis of the Vibrio fischeri autoinducer, a signal involved in the cell density-dependent activation of bioluminescence, is directed by the luxI gene product. The LuxI protein catalyzes the synthesis of N-acyl-homoserine lactones from S-adenosylmethionine and acylated-acyl carrier protein. We have gained an appreciation of the LuxI regions and amino acid residues involved in autoinducer synthesis by isolating and analyzing mutations generated by random and site-specific mutagenesis of luxI. By random mutagenesis we isolated 13 different single amino acid substitutions in the LuxI polypeptide. Eleven of these substitutions resulted in no detectable autoinducer synthase activity, while the remaining two amino acid substitutions resulted in reduced but detectable activity. The substitutions that resulted in no detectable autoinducer synthase activity mapped to two small regions of LuxI. In Escherichia coli, wild-type luxI showed dominance over all of the mutations. Because autoinducer synthesis has been proposed to involve formation of a covalent bond between an acyl group and an active-site cysteine, we constructed site-directed mutations that altered each of the three cysteine residues in LuxI. All of the cysteine mutants retained substantial activity as an autoinducer synthase in E. coli. Based on the analysis of random mutations we propose a model in which there are two critical regions of LuxI, at least one of which is an intimate part of an active site, and based on the analysis of site-directed mutations we conclude that an active-site cysteine is not essential for autoinducer synthase activity.

LuxR is required for cell density-dependent activation of the Vibrio fischeri luminescence (lux) genes. It has not been possible to study full-length LuxR in vitro, but a polypeptide containing the C-terminal transcriptional-activator domain of LuxR (LuxRdeltaN) has been purified, and its binding to lux regulatory DNA has been investigated. By itself, LuxRdeltaN interacts with a region of lux regulatory DNA that is upstream of the lux box, which is a 20-bp element that is required for LuxR activation of the luminescence operon. Individually, neither the purified LuxRdeltaN nor RNA polymerase binds to the lux box region, but together the two proteins bind in synergy to the lux box-luxI promoter region. We show that binding of LuxRdeltaN to the upstream region is not a prerequisite for its synergistic binding with RNA polymerase to the lux box and the luxI promoter region. We also show that LuxRdeltaN and RNA polymerase are both required and sufficient for transcriptional activation of the lux operon. This argues against the hypothesis that LuxR functions to alleviate repression of the lux operon by another cellular factor. Rather, our data support the view that LuxR functions as an accessory factor that enables RNA polymerase to bind to and initiate transcription from the promoter of the lux operon.

Many bacteria use acyl homoserine lactone signals to monitor cell density in a type of gene regulation termed quorum sensing and response. Synthesis of these signals is directed by homologs of the luxi gene of Vibrio fischeri. This communication resolves two critical issues concerning the synthesis of the V. fischeri signal. (i) The luxI product is directly involved in signal synthesis-the protein is an acyl homoserine lactone synthase; and (ii) the substrates for acyl homoserine lactone synthesis are not amino acids from biosynthetic pathways or fatty acid degradation products, but rather they are S-adenosylmethionine (SAM) and an acylated acyl carrier protein (ACP) from the fatty acid biosynthesis pathway. We purified a maltose binding protein-LuxI fusion polypeptide and showed that, when provided with the appropriate substrates, it catalyzes the synthesis of an acyl homoserine lactone. In V. fischeri, luxi directs the synthesis of N-(3-oxohexanoyl) homoserine lactone and hexanoyl homoserine lactone. The purified maltose binding protein-LuxI fusion protein catalyzes the synthesis of hexanoyl homoserine lactone from hexanoyl-ACP and SAM. There is a high level of specificity for hexanoyl-ACP over ACPs with differing acyl group lengths, and hexanoyl homoserine lactone was not synthesized when SAM was replaced with other amino acids, such as methionine, S-adenosylhomocysteine, homoserine, or homoserine lactone, or when hexanoyl-SAM was provided as the substrate. This provides direct evidence that the LuxI protein is an auto-inducer synthase that catalyzes the formation of an amide bond between SAM and a fatty acyl-ACP and then catalyzes the formation of the acyl homoserine lactone from the acyl-SAM intermediate.

Synthesis of the autoinducer signal involved in the cell density-dependent activation of Vibrio fischeri luminescence is directed by luxI. The autoinducer is N-(3-oxohexanoyl)homoserine lactone, and little is known about its synthesis. We have measured autoinducer synthesis by amino acid auxotrophs of Escherichia coli that contained luxI on a high-copy-number plasmid. Experiments with cell suspensions starved for methionine or homoserine show that either methionine or S-adenosylmethionine but not homoserine or homoserine lactone is required for autoinducer synthesis. The S-adenosylmethionine synthesis inhibitor cycloleucine blocks methionine-dependent autoinducer synthesis. Thus, it appears that S-adenosylmethionine rather than methionine is the molecule required for autoinducer synthesis. The amount of 15N-labeled methionine incorporated into the autoinducer by growing cultures of a homoserine and a methionine auxotroph was measured by mass spectrometry. The labeling studies show that even in the presence of homoserine, almost all of the autoinducer produced contains the 15N label from methionine. Thus, it appears that S-adenosylmethionine serves as the amino acid substrate in the luxI-dependent synthesis of the V. fischeri autoinducer.

The Vibrio fischeri luminescence genes are activated by the transcription factor LuxR in combination with a diffusible signal compound, N-(3-oxohexanoyl) homoserine lactone, termed the autoinducer. We have synthesized a set of autoinducer analogs. Many analogs with alterations in the acyl side chain showed evidence of binding to LuxR. Some appeared to bind with an affinity similar to that of the autoinducer, but none showed a higher affinity, and many did not bind as tightly as the autoinducer. For the most part, compounds with substitutions in the homoserine lactone ring did not show evidence of binding to LuxR. The exceptions were compounds with a homocysteine thiolactone ring in place of the homoserine lactone ring. Many but not all of the analogs showing evidence of LuxR binding had some ability to activate the luminescence genes. None were as active as the autoinducer. While most showed little ability to induce luminescence, a few analogs with rather conservative substitutions had appreciable activity. Under the conditions we employed, some of the analogs showing little or no ability to induce luminescence were inhibitors of the autoinducer.

The importance of accurate demographic information is reflected in the United States Constitution, Article 1, which provides for a decennial census of this country's human population. Bacteria also conduct a census of their population and do so more frequently, more efficiently, and as far we know, with little if any of the political contentiousness caused by human demographers. Many examples have been found of particular bacterial genes, operons, or regulons that are expressed preferentially at high cell densities. Many of these are regulated by proteins related to the LuxR and LuxI proteins of Vibrio fischeri, and by a diffusible pheromone called an autoinducer. LuxR and LuxI and their cognate autoinducer (3-oxohexanoyl homoserine lactone, designated VAI-1) provide an important model to describe the functions of this family of proteins. LuxR is a VAI-1 receptor and a VAI-1-dependent transcriptional activator, and LuxI directs the synthesis of VAI-1. VAI-1 diffuses across the bacterial envelope, and intracellular concentrations of it are therefore strongly increased by nearby VAI-1-producing bacteria. Similar systems regulate pathogenesis factors in Pseudomonas aeruginosa and Erwinia spp., as well as T1 plasmid conjugal transfer in Agrobacterium tumefaciens, and many other genes in numerous genera of gram-negative bacteria. Genetic analyses of these systems have revealed a high degree of functional conservation, while also uncovering features that are unique to each.

The Vibrio fischeri luminescence genes are regulated by the LuxR protein and an N-acyl homoserine lactone compound termed the autoinducer. The C-terminal one-third of LuxR contains a domain that can interact with the transcription complex and activate the luminescence genes. On the basis of limited evidence it has been suggested that the N-terminal two-thirds of LuxR constitutes a domain that serves to bind the autoinducer. We show that tritium-labeled autoinducer binds to Escherichia coli cells in which LuxR is overexpressed. We also show that tritium-labeled autoinducer binds to E. coli in which truncated LuxR proteins missing portions of the C-terminal domain are expressed but does not bind to E. coli cells in which truncated LuxR proteins missing portions of the N-terminal region are expressed. Our results provide evidence that the autoinducer binds to LuxR and that in E. coli the N-terminal two-thirds of LuxR can fold into a polypeptide capable of binding the autoinducer in the absence of the C-terminal domain.

LuxR, the Vibrio fischeri luminescence gene (lux) activator, is the best-studied member of a family of bacterial transcription factors required for cell density-dependent expression of specific genes involved in associations with eukaryotic hosts. Neither LuxR nor any other LuxR homolog has been shown to bind DNA directly. We have purified the LuxR C-terminal transcriptional activator domain from extracts of recombinant Escherichia coli in which this polypeptide was expressed. The purified polypeptide by itself binds to lux regulatory DNA upstream of the lux box, a 20-bp palindrome that is required for LuxR activity in vivo, but it does not bind to the lux box. However, the LuxR C-terminal domain together with RNA polymerase protects a region including the lux box and the lux operon promoter from DNase I cleavage. There is very little protection of the lux operon promoter region from DNase I digestion in the presence of RNA polymerase alone. Apparently, there is a synergistic binding of the LuxR C-terminal domain and RNA polymerase to the promoter region. The upstream binding region for the purified polypeptide encompasses a binding site for cAMP receptor protein (CRP). Under some conditions, CRP binding can block the binding of the LuxR C-terminal domain to the upstream binding region, and it can also block the synergistic binding of the LuxR C-terminal domain and RNA polymerase to the lux box and luminescence gene promoter region. This description of DNA binding by the LuxR C-terminal domain should lead to an understanding of the molecular interactions of the LuxR family of transcriptional activators with regulatory DNA.

Autoinduction is a conserved mechanism of cell density-dependent gene regulation that occurs in a variety of gram-negative bacteria. Autoinducible luminescence in Vibrio fischeri requires a transcriptional activator, LuxR, while a LuxR homolog, LasR, activates elastase expression in Pseudomonas aeruginosa. Both LuxR and LasR require specific signal molecules, called autoinducers, for activity. We show here the activation in Escherichia coli of the V. fischeri luminescence (lux) operon by LasR and of the P. aeruginosa elastase gene (lasB) by LuxR when each is in the presence of its cognate autoinducer. Neither LuxR nor LasR showed appreciable activity with the heterologous V. fischeri or P. aeruginosa autoinducer. This supports the view that there is a direct interaction of each transcriptional activator with its proper autoinducer and suggests that there are conserved, autoinduction-related elements within the promoter regions of these genes.

The Vibrio fischeri luminescence (lux) genes are activated at sufficiently high culture densities by the transcriptional activator LuxR in combination with a diffusible signal compound termed autoinducer. We have used antibodies directed against LuxR in immunoprecipitation experiments to study the subcellular location of this transcription factor. The LuxR polypeptide was detected in membranes and not in the soluble pool of cytoplasmic proteins from V. fischeri. LuxR was not released from the membranes by 0.6 M KCl or by the nonionic detergents Nonidet P-40, N-octyl-beta-D-glucopyranoside, and Triton X-100. LuxR and a number of other V. fischeri proteins were released from the membranes by EDTA. The autoinducer had no detectable influence on the subcellular location of LuxR. In spheroplasts, neither the abundance nor the molecular mass of the LuxR antigen was influenced by treatment with proteinase K. Together with other information, these results indicate that LuxR is an amphipathic protein that is associated with the cytoplasmic membrane of V. fischeri.

In Escherichia coli, transcription of the inducible Vibrio fischeri luminescence operon, luxICDABE, has been reported to require sigma 32, the product of rpoH. Consistent with previous studies, we report that an E. coli delta rpoH mutant, KY1601 containing luxICDABE and luxR, which codes for the activator of luxICDABE transcription on a plasmid (pJE202), was weakly luminescent. Transformation of this E. coli strain with a plasmid containing rpoH under the control of the tac promoter resulted in high levels of cellular luminescence. However, the level of expression of the pJE202 luxICDABE was also high in E. coli 1603, a delta rpoH mutant with a second-site mutation that resulted in sigma 32-independent overexpression of the groE operon. Apparently, sigma 32 is not directly required for the transcription of luxICDABE in E. coli but is required for sufficient expression of groE, which is in turn required for the transcription of luxICDABE. This conclusion is supported by the finding that E. coli groE mutants containing pJE202 were weakly luminescent. In the E. coli delta rpoH mutant KY1601, the sigma 32 requirement for the transcription of luxICDABE was partially compensated for by the addition of saturating concentrations of the inducer to the culture medium and largely compensated for when cells were transformed with a luxR overexpression vector. These data support the hypothesis that sigma 32 is not required for transcription of luxICDABE. Rather, it appears that the products of groE are required for the folding of LuxR into an active protein, like they are for the folding of several other proteins.

Vibrio fischeri ES114 is an isolate representing the specific bacterial light organ symbiont of the squid Euprymna scolopes. An interesting feature of this strain of V. fischeri is that it is visibly luminous within the light organ of the squid host but is nonluminous when grown under standard laboratory conditions. Luminescence can be restored in laboratory culture, however, by the addition of autoinducer, a species-specific inducer of the V. fischeri luminescence (lux) genes. Most other isolates of V. fischeri produce autoinducer in sufficient quantities to induce luminescence in laboratory culture. We have cloned an 8.8-kb DNA fragment from V. fischeri ES114 that encodes all of the functions necessary for luminescence in Escherichia coli in the absence of exogenous autoinducer. This DNA contains both of the recognized V. fischeri lux regulatory genes, one of which (luxI) directs E. coli to synthesize autoinducer. The organization of the individual lux genes within this DNA fragment appears to be the same as that in the other strains of V. fischeri studied; the restriction map of the V. fischeri ES114 lux DNA has diverged substantially, however, from the largely conserved maps of V. fischeri MJ1 and ATCC 7744. Although E. coli containing the V. fischeri ES114 lux DNA synthesizes considerable amounts of autoinducer, V. fischeri ES114 synthesizes autoinducer only in small amounts, even when transcription of the lux genes, including luxI, is activated by the addition of exogenous autoinducer. Nonetheless, transconjugants of V. fischeri ES114 that contain multicopy plasmids bearing the ES114 lux genes synthesize sufficient autoinducer to induce luminescence. These results suggest that V. fischeri ES11r does not lack a functional luxl, nor is it deficient in the ability to synthesize metabolic precursors for autoinducer synthesis.

The Vibrio fischeri luminescence (lux) genes are regulated by the 250-amino-acid-residue LuxR protein and a V. fischeri metabolite termed autoinducer. The V. fischeri lux regulon consists of two divergently transcribed units. Autoinducer and LuxR activate transcription of the luxICDABE operon and autoregulate the luxR transcriptional unit. LuxR proteins with C-terminal truncations of up to 40 amino acid residues coded by plasmids with luxR 3'-deletion mutations are functional in negative autoregulation as demonstrated by using a luxR::lacZ transcriptional fusion as a luxR promoter probe in Escherichia coli. The truncated LuxR proteins showed little or no ability to activate transcription of luxICDABE, as indicated by using luminescence as a sensitive indicator of promoter strength in E. coli. Besides having no detectable activity as positive regulators of luxICDABE, LuxR proteins with C-terminal truncations of more than 40 amino acid residues had reduced or no detectable activity as negative autoregulators. The results suggest that amino acid residues in LuxR prior to no. 211 are sufficient for lux DNA binding. Residues in the region of 211 to 250 constitute a C-terminal tail that appears to be involved in activation of luxICDABE transcription either by interacting physically with the transcription initiation complex or by affecting lux DNA in the vicinity of the promoter.

The Vibrio fischeri luminescence genes are activated by an autoinducer and the 250-amino acid residue LuxR protein. To develop a general view of LuxR structure and function, a set of luxR 5'-deletion mutations was generated. Ten luxR mutant plasmids encoding active LuxR proteins with deletions ranging from residues 2-5 (delta 2-5) to residues 2-182 (delta 2-182) were studied. The degree of transcriptional activation of luminescence genes by the truncated LuxR proteins ranged from 0.01% to greater than 200% of the wild-type level. LuxR proteins with small deletions (up to delta 2-20) were active and remained autoinducer-dependent, LuxR proteins with deletions between residues 2-58 and 2-138 showed low activity and were not affected by autoinducer, and LuxR proteins with large deletions such as the delta 2-162 protein were highly active and autoinducer-independent. However, proteins with deletions equal to or greater than delta 2-20 were unable to autoregulate luxR. Our data indicate there is a C-terminal LuxR domain capable of functioning as a transcriptional activator. We suggest that an N-terminal region of LuxR starting between residues 20 and 58 and extending to the region of residues 138-162 masks the activator function of the C-terminal domain. Residues prior to position 20 are needed for autoregulatory function. Experiments showing that wild-type luxR is dominant over luxR genes coding the delta 2-58 through delta 2-138 proteins indicate the N-terminal arm masks lux DNA binding.

Expression of Vibrio fischeri luminescence genes requires an inducer, termed autoinducer, and a positive regulatory element, the luxR gene product. A plasmid containing a tac promoter-controlled luxR was mutagenized in vitro with hydroxylamine, and luxR mutant plasmids were identified by their inability to complement a luxR deletion mutation in trans. Sixteen luxR mutant plasmids were obtained, ten of which encoded full-length but inactive luxR gene products as demonstrated by a Western immunoblot analysis. The effects of 1 of the 10 mutations could be overcome by the addition of autoinducer at a high concentration. The mutations in each of the 10 mutant plasmids that directed the synthesis of an inactive LuxR protein were identified by DNA sequencing. Of the 10 proteins encoded by the mutant luxR plasmids, 9 differed from the normally active LuxR in only a single amino acid residue. The amino acid residue substitutions in the proteins encoded by the nine mutant luxR genes clustered in two regions. One region around the middle of the polypeptide encoded by luxR was hypothesized to represent an autoinducer-binding domain, and the other region towards the carboxy terminus of the gene product was hypothesized to constitute a lux operator DNA-binding domain or a lux operator DNA recognition domain.

The enzymes for luminescence in Vibrio fischeri are induced by the accumulation of a species-specific metabolite (autoinducer) in the culture medium. Tritium-labeled autoinducer was used to study the mechanism of autoinduction. When 3H-autoinducer was added to suspensions of V. fischeri or Escherichia coli, cellular concentrations equaled external concentrations. For V. fischeri, equilibration of 3H-autoinducer was rapid (within 20 s), and greater than 90% of the cellular tritium remained in unmodified autoinducer. When V. fischeri or E. coli cells containing 3H-autoinducer were transferred to autoinducer-free buffer, 85 to 99.5% of the radiotracer escaped from the cells, depending on the strain. Concentrations of autoinducer as low as 10 nM, which is equivalent to 1 or 2 molecules per cell, were sufficient for induction, and the maximal response to autoinducer occurred at about 200 nM. If external autoinducer concentrations were decreased to below 10 nM after induction had commenced, the induction response did not continue. Based on this study, a model for autoinduction is described wherein autoinducer association with cells is by simple diffusion and binding of autoinducer to its active site is reversible.

Expression of the Vibrio fischeri luminescence genes (lux genes) requires two transcriptional activators: the V. fischeri luxR gene product with autoinducer and the cyclic AMP (cAMP) receptor protein (CRP) with cAMP. It has been established that autoinducer and the luxR gene product are required for transcriptional activation of the luxICDABE operon, which contains a gene required for autoinducer synthesis and genes required for light emission. However, the role of cAMP-CRP in the induction of luminescence is not clear. We examined transcriptional control of the lux genes in Escherichia coli, using catabolite repression mutants carrying lux DNA-containing plasmids. Transcriptional fusions between the lacZ gene on Mu dI and luxR were used to assess luxR promoter activity, and the luxAB genes (which encode the two luciferase subunits) were used as a natural reporter of luxICDABE promoter activity. A plasmid containing luxR under control of the cAMP-CRP-independent tac promoter was constructed to direct the synthesis of the luxR gene product in cells containing compatible luxR::Mu dI insertion mutant plasmids. In E. coli, cAMP-CRP activated transcription of luxR and concurrently decreased luxICDABE transcription. In the presence of relatively high levels of the luxR gene product, cAMP and CRP were not required for induction of the luxICDABE operon. The luxR gene product in the presence of autoinducer activated transcription of the luxICDABE operon, as has been shown previously, and we demonstrate that it also decreased luxR transcription. Apparently, control of the V. fischeri luminescence genes involves a regulatory circuit in which cAMP and CRP activate luxR transcription and in turn the luxR gene product activates transcription of the operon responsible for light emission (uxICDABE). Furthermore, in lux gene regulation cAMP-CRP and autoinducer-LuxR protein appear to function as transcriptional antagonists.

Under certain conditions glucose represses the autoinducible synthesis of luminescence enzymes in Vibrio fischeri. To examine the genetic regulation of luminescence more closely, Escherichia coli catabolite repression mutants were transformed with a plasmid (pJE202) that contains V. fischeri genes specifying the luminescence enzymes and encoding regulatory functions for luminescence (the lux genes) or with plasmids (pJE413 and pJE455) containing transcriptional fusions between the lacZ gene on transposon mini-Mu and specific genes in each of the two lux operons. Unless cyclic AMP (cAMP) was added to the growth medium, an adenylate cyclase deletion mutant containing pJE202 produced very little light and low levels of the light-emitting enzyme luciferase. When grown in the presence or absence of cAMP, a cAMP receptor protein (CRP) deletion mutant produced low levels of light and luciferase. A mutant that does not make cAMP but does make an altered CRP which does not require cAMP for activity produced induced levels of luminescence after transformation with pJE202. To test the effects of cAMP and CRP on each of the two lux operons separately rather than on both together, the E. coli catabolite repression mutants were transformed with pJE413 and pJE455. From measurements of beta-galactosidase and luciferase activities it appeared that cAMP and CRP affected transcription of both lux operons. In the presence of autoinducer and its receptor, transcription of the operon encoding all of the luminescence genes except the receptor gene appeared to be activated by cAMP and CRP, whereas in the absence of the receptor, cAMP and CRP appeared to decrease transcription of this operon. Transcription of the operon encoding the autoinducer receptor appeared to be stimulated by cAMP and CRP in the absence of the receptor itself. These results demonstrate that cAMP and CRP are required for proper control of the V. fischeri luminescence system and suggest that lux gene transcription is required by a complex mechanism.