*Corresponding author. Mailing address: Institute of Infections and Immunity, University of Nottingham, Queen's Medical Centre, Nottingham NG7 2UH, UK. Tel.: +44 (115) 970 9907; Fax: +44 (115) 970 9923; E-mail: email@example.com
The luxCDABE operon of Photorhabdus luminescens has been cloned and engineered as an easily mobilisable cassette flanked by sites for commonly used restriction enzymes. Constitutive and promoter probe plasmids utilising the P. luminescens luxCDABE have been constructed using a number of compatible replicons and antibiotic markers. Complementary to these plasmids, a range of promoterless and constitutive luxCDABE mini-Tn5 derivatives has been constructed. The potential of coupling mini-Tn5 luxCDABE promoter probe transposons with automated luminometry and photometry to screen for mutants that exhibit growth phase variation in gene expression is demonstrated.
A wide range of biological sensors has been used to monitor gene expression (see [1, 2]). The application of these biosensors often involves an endpoint determination of the activity of a reporter enzyme, with samples taken at intervals for evaluation of activity in assays requiring additional reagents. Although these methods are often sensitive and reasonably rapid, they are typically invasive and destructive and have to be performed off-line. As such, they do not act as real time reporters.
Bioluminescent reporters offer the facility to quantify gene expression at high sensitivity over a large dynamic range in real time and non-destructively. Whereas many reporter systems suffer from the disadvantage of high background expression due to the activity of host enzymes, with bioluminescence this is much less of a problem as naturally bioluminescent bacteria are extremely rare outside the marine environment.
The biochemistry and genetics of bacterial bioluminescence, and the range of applications for investigating microbial physiology (including the monitoring of both cellular viability and gene expression), have been extensively reviewed (for example [3, 4]). The lux genes essential for luminescence are arranged in a single operon, luxCDABE. luxCDE encode a fatty acid reductase complex involved in synthesis of the fatty aldehyde substrate for the luminescence reaction catalysed by the luciferase LuxAB subunits . Two of the substrates for the bioluminescent reaction, reduced flavin mononucleotide (FMNH2) and molecular oxygen, are both readily available in aerobic bacteria. Biosensors which employ the luciferase coded by the luxAB genes alone are frequently used both in Gram-negative and Gram-positive bacteria [5–7]; however, the necessity to supply exogenous aldehyde is a drawback for on-line analysis or measurements in planta or in vivo in experimental animal models. Utilisation of the luxCDABE operon in Gram-negative bacteria overcomes this aldehyde requirement .
Although all of the lux genes so far isolated are derived from Gram-negative bacteria , their functional properties can vary between species. The particular source of the lux genes to be employed as a reporter is therefore an important consideration. The limited temperature range of the most commonly utilised lux genes, derived from Vibrio fischeri (<30°C) and Vibrio harveyi (<37°C), can be restrictive in some applications; e.g., in the study of gene expression in animal pathogens. In contrast, the enzymes encoded by the luxCDABE operon of the terrestrial bacterium Photorhabdus (Xenorhabdus) luminescens are functional at temperatures as high as 45°C [5, 8]. It was therefore envisaged that sensors based on the P. luminescens luxCDABE operon would permit greater flexibility and ease of use in Gram-negative bacteria than the available luxAB or luxCDABE systems derived from V. fischeri or V. harveyi. To this end we have constructed a range of plasmid and mini-Tn5 vectors containing an engineered luxCDABE cassette derived from the P. luminescens lux operon and designed for use in Gram-negative bacteria.
Promoter probe transposons insert promoterless reporter genes randomly into the chromosome, often creating transcriptional fusions and insertional mutations. Using agar plate techniques the effect of given chemicals and treatments, such as temperature shock, can be used to detect gross changes in expression by looking at differences over a period of time . In addition to the applications associated with the original constructs of de Lorenzo et al. , expansion of the available mini-Tn5 range to include promoterless luxCDABE cassettes now provides the opportunity to screen throughout growth for genes expressed at a particular phase of growth, or in response to a specific physical, chemical or culture condition. The use of luciferases to report on gene expression throughout the growth of a bacterial culture is, however, a classically labour intensive process with a low throughput. In this study we have used an automated photometer/luminometer (Lucy1, Anthos Labtech, Salzburg, Austria) to perform these experiments in a high throughput, microplate format.
2Materials and methods
2.1Bacterial strains, growth conditions and conjugation
Bacterial strains used in this study are detailed in Table 1. Escherichia coli JM109 was used as the host for plasmids not requiring a λpir host and E. coli CC118 λpir was used as the host for plasmids requiring a λpir host. The pUT mini-Tn5 transfer plasmid is only maintained in a λpir host and can be mobilised from E. coli S17-1 λpir into the target species of interest by conjugation. For conjugation the donor and recipient strains were grown overnight in 5 ml L-broth, Lennox (Difco) (LB) at the appropriate temperature (37°C for E. coli/pUT mini-Tn5 luxCDABE Km2 and 30°C for Aeromonas hydrophila and Chromobacterium violaceum) with antibiotic. Harvested cells were gently washed twice in LB and resuspended in 1/10th of the original culture volume. 100-μl volumes of donor and recipient cells were gently mixed, spot-inoculated onto the surface of LB agar plates (Oxoid No. 1 agar at 1.5% (w/v)) and incubated at 30°C for 4–8 h. The cells were scraped from the surface and washed once with 1 ml LB. Dilutions were spread onto Aeromonas selective medium (Difco), or LB, agar plates containing appropriate antibiotics and incubated for 48 h at 30°C. Bioluminescent transconjugants were purified by sub-culturing twice on selective medium.
Genomic DNA was purified according to a standard protocol , plasmid DNAs were isolated by alkaline lysis  and further purified using caesium chloride gradients  or Qiagen plasmid preparation columns. Restriction enzyme digests and DNA ligations were performed according to the manufacturers' instructions (Promega, Boehringer Mannheim, or Pharmacia). PCR amplifications were performed according to a standard protocol  and the sequence of PCR-derived DNA was checked for polymerase errors. Oligonucleotides were synthesised by the Biopolymer Synthesis and Analysis Unit, University of Nottingham. DNA sequencing was performed by the University of Nottingham Automated Sequencing Facility.
2.3Luminometry and photometry
Bioluminescent colonies on agar plates were imaged using either an Argus-100 photon-imaging camera (Hamamatsu Photonics) or a Luminograph LB980 photon video camera (E.G. and G. Berthold). For automated luminometry and photometry (using absorbance at 405 nm as a measure of optical density), overnight cultures in LB were diluted 1:100 in fresh medium (LB or Minimal M9 ) and 0.2 ml was used to inoculate each microplate well. Where stated, HPLC-purified, synthetic N-hexanoyl-l-homoserine lactone (HHL) [10, 11] was added to a final concentration of 1 μM from a 10-mM stock solution in acetonitrile (far UV grade). Assays were performed with incubation at 30°C and shaking to ensure a homogeneous distribution of cells for readings, rather than the provision of vigorous aeration. Data were written to a spreadsheet format compatible with Microsoft Excel for further analysis. In each case the optical density data were written in milli-absorbance units (OD405nm). Plots of log10 (luminescence/optical density) against optical density are used to represent the data. Results show the variation of bioluminescence per bacterium with increasing population size. Hence constitutive promoter::lux fusions show a gradient of zero and induced promoter::lux fusions show a positive gradient. The point in the growth phase where induction takes place can therefore be identified as the point where the gradient begins to increase. All the plots shown are representative of at least three replicates.
3Results and discussion
3.1Cloning of the Photorhabdus luminescens ATCC29999 lux operon and construction of a luxCDABE cassette
Shotgun cloning of Photorhabdus luminescens ATCC29999 (Hb strain) genomic DNA into plasmid vectors pACYC184 , pT7T3 18U (Pharmacia) and pUC19  yielded highly bioluminescent E. coli JM109 clones containing plasmids designated pSB311 (in pACYC184), pSB312 (in pT7T3 18U with lux transcribed in the opposite direction to Plac), pSB313 (in pT7T3 18U with lux transcribed in the same direction as Plac) and pSB341 (in pUC19). Each contained the P. luminescens lux operon on a 6.96-kb EcoRI fragment (Fig. 1A). The P. luminescens ATCC29999 lux operon (GenBank accession number M90093) contains no sites for cleavage by commonly used multicloning site restriction enzymes (EcoRI, BamHI, PstI, SalI, SmaI, KpnI and XhoI) . The 6.96-kb EcoRI fragment contains approximately 450 bp of sequence not required for expression of luxCDABE upstream of luxC and 700 bp downstream of luxE. To construct a luxCDABE reporter cassette especially suited to forming transcriptional fusions, this fragment was reduced in size to contain only DNA coding for the luxCDABE genes essential for bioluminescence in the absence of exogenous aldehyde.
For modification of the luxC upstream region a 2.4-kb XbaI fragment containing luxC and part of luxD was deleted from pSB341 to create pSB342. The 5.5-kb XbaI fragment containing luxABE and part of luxD was deleted from pSB313 to create pSB313X. A primer based on the known sequence at the start of the luxC gene (5′-CCCGGGTCTAGAATTCAGGCTTGGAGGATACGTATGACTAAAAAAATTTC-3′)  was designed to introduce SmaI, XbaI, EcoRI and SnaBI restriction sites and a synthetic ribosome binding site upstream of the luxC ATG codon (underlined) when used in a PCR amplification with an M13 universal primer (5′-GTAAAACGACGGCCAGT-3′) and pSB313X DNA template. Translational stop codons in non-coding frames are present downstream of the luxC initiation codon at nucleotides 2–4 (TGA) and 6–8 (TAA). The resultant 2.0-kb PCR product was digested with XbaI and cloned into the XbaI site of pSB342 to produce pSB373.
For modification of the region downstream of luxE an EcoRV site conveniently positioned at the end of this gene was chosen for the terminus of the lux cassette fragment. As another EcoRV site is present in luxC (Fig. 1A), a 2.1-kb HpaI fragment from pSB312 containing the whole of luxE and part of luxB was cloned into the SmaI site of pT7T3 18U to create pSB380. The region downstream from luxE in this plasmid was then deleted between the EcoRV site at the end of luxE and the vector HincII site to create pSB381. An EcoRI fragment containing the PCR-modified promoterless lux genes from pSB373 was cloned into the unique EcoRI site of pSB381 creating pSB386 and the sequence between the two BglII sites (in luxE) deleted to produce pSB387, a plasmid containing a 5.8-kb EcoRI-PstI fragment with only the luxCDABE coding region and the modified luxC upstream region.
To incorporate convenient restriction sites at both ends of the luxCDABE cassette, the 1.2-kb EcoRI Kmr fragment from pUC4K (Pharmacia) was introduced into the unique EcoRI site on pSB387 to provide a PstI site upstream of luxC (pSB388). The 5.8-kb PstI fragment containing luxCDABE was then cloned into pUC4K in place of the Kmr cassette to yield pSB389. The lux cassette in pSB389 is flanked at both ends by PstI, SalI, BamHI and EcoRI sites (Fig. 1B).
3.2Construction of promoter probe plasmids
A set of promoter probe vectors was designed to encompass a range of antibiotic resistance markers and origins of replication. For a low copy number pSB322 derivative, conferring ampicillin resistance and containing a single EcoRI site upstream of luxC suitable for promoter cloning, pSB377 (Fig. 2A) was constructed. A 5.5-kb EcoRV fragment from pSB312 was cloned into the HpaI site of pHG177  to produce pSB375. The 2-kb EcoRI-XbaI fragment from pSB373 (containing the modified region upstream of luxC, luxC and part of luxD) was then used to replace the region between the vector EcoRI site and the XbaI site in luxD and create the bioluminescent promoter probe plasmid pSB377. For a medium copy number pACYC184 derivative, conferring tetracycline resistance and containing a single EcoRI site upstream of luxC suitable for promoter cloning, pSB384 (Fig. 2B) was constructed. A 5.5-kb EcoRV fragment from pSB312 was cloned into pACYC184 in place of a 400-bp PvuII fragment upstream of the chloramphenicol resistance gene to create pSB382. The 5.8-kb EcoRI fragment from pSB373 then was cloned into the EcoRI site internal to the pSB382 cat gene to create pSB383. The region between the two BglII sites was then deleted to form pSB384.
For pSB395 (Fig. 2C), a medium copy number, broad host range mobilisable plasmid (oriV, oriT) which confers Tcr and contains a single EcoRI site upstream of luxC suitable for promoter cloning, a 5.8-kb luxCDABE EcoRI-PstI fragment from pSB389 was cloned between the EcoRI-PstI sites of pRK415 .
For useful control vectors and plasmids for tagging cells in viability studies the plasmids pSB390 (5.8-kb luxCDABE PstI cassette from pSB389 cloned into PstI site introduced into pACYC184 by cloning PstI flanked kanamycin cassette from pUC4K as an EcoRI fragment, lux transcription is from Pcat), pSB416 (luxCDABE cassette on a 5.8-kb PstI fragment from pSB390 cloned into pUC18Not, ), pSB417 (luxCDABE cassette on a 5.8-kb PstI fragment from pSB390 cloned into pUC18Sfi with lux transcription from Plac) and pSB418 (luxCDABE cassette on 5.8-kb PstI fragment from pSB390 cloned into pUC18Sfi, with lux in the opposite orientation to pSB417) were constructed. These constructs confer a highly bioluminescent host phenotype through transcriptional read-through from vector promoters and are a convenient source of the promoterless luxCDABE cassette flanked by different restriction sites. Constitutive bioluminescence expression from recombinant organisms carrying any of these plasmids are applicable to studies requiring an evaluation of cellular viability, e.g., in response to an inimical process and in tracking studies (see [6, 7, 17]). Inducible bioluminescence from promoter probe plasmids, constructed by either shotgun or rational cloning of promoter elements, can be used for the monitoring of conditional gene expression in response to controlled prevailing conditions. The absence of an upstream, in frame, stop codon should be noted here, as it is conceivable that a fusion protein may be produced, which could have a significant effect upon the resultant bioluminescence. The choice of replicon and antibiotic resistance marker in these plasmids offers flexibility with respect to the host range in Gram-negative bacteria and the incompatibility class of additional plasmids present in trans.
3.3Construction of mini-Tn5 luxCDABE constructs
The mini-Tn5 transposons of de Lorenzo and co-workers are established tools for the investigation of the molecular biology of Gram-negative bacteria [9, 16] and include lacZ, luxAB (V. harveyi), phoA and xylE promoter probe reporters . Promoterless mini-Tn5 luxCDABE constructs conferring Cmr, Kmr, and Sm/Spr (Fig. 3 A–C) were constructed by cloning a luxCDABE NotI cassette from pSB416 into the unique NotI site of the corresponding pUTmini-Tn5 suicide plasmids. In the case of mini-Tn5 luxCDABE Tc, the promoterless luxCDABE cassette replaced luxAB in mini-Tn5 luxAB Tc  (Fig. 3D). A mini-Tn5 derivative conferring constitutive lux expression caused by read-through from the internal Kmr gene promoter was made by insertion of the luxCDABE NotI cassette into pUTmini-Tn5 Km1  (Fig. 3E). The mini-Tn5 luxCDABE reporter constructs are particularly suited to tagging cells for tracking studies as they are not self-transposable and, unlike many plasmid reporters, are stably maintained in the absence of antibiotic. Furthermore, in the construction of the mini-Tn5 luxCDABE promoter probe transposons an in frame stop codon (TAG) was introduced 18 codons upstream of the luxC ATG.
Additionally, a plasmid useful for creating marker-exchange mutants and investigating transcription of chromosomally located genes through monitoring of bioluminescence (Fig. 3F) was constructed. The EcoRI fragment containing luxCDABE Km2 from pUTmini-Tn5 luxCDABE Km2 was cloned into pUC4K in place of the Kmr cassette and the NotI site downstream from luxE was removed by deletion of a small PstI fragment to produce pSB420. This modified EcoRI cassette was then cloned into pUC18Not to create pSB421 containing a NotI cassette with the luxCDABE Km2 element flanked by EcoRI and KpnI sites. Chromosomal integration of this reporter element can be achieved by recombination between chromosomal sequences and the homologous DNA cloned between the KpnI sites downstream of the kanamycin resistance gene and into the EcoRI site upstream of luxC. The NotI cassette is suited to this purpose because it is unlikely that sequences used for homologous recombination will contain NotI restriction sites.
3.4Screening transposon promoter probe libraries for growth phase and medium dependent promoters
The bacterial phenotype is a dynamic phenomenon, which is modulated in response to the changing environment. The effects of a growing population of bacteria on this environment provide important parameters controlling phenotypic changes at the level of promoter activity, e.g., in quorum sensing [10, 11, 18, 19] and the entry into stationary phase . The study of the regulation of these genes depends upon the ability to identify the time and conditions inducing gene expression, a use to which the luxCDABE operon as a reporter is ideally suited.
In quorum sensing, gene expression at high cell density is induced after the accumulation of a signal molecule in the extracellular medium of the bacterial population . In many Gram-negative bacteria the signal molecule is an N-acylhomoserine lactone (AHL) . Using sensor strains for AHL signal molecules, it has been possible to screen gene libraries for a number of Gram-negative bacteria and identify luxI homologue clones (e.g., [11, 19]). Marker exchange and transposon mutagenesis strategies have allowed the creation of mutants in the luxI homologue gene and the subsequent screening for associated phenotypes (e.g., [10, 19]). In cases where either no obvious phenotype is associated with the knockout of quorum sensing or the identification of further traits regulated by quorum sensing is desired, screening randomly generated promoterless transposon libraries in the presence and absence of the cognate AHL may detect regulated genes. We identified this type of analysis as one potentially suited to the use of a mini-Tn5 luxCDABE construct coupled with analysis by automated luminometry and photometry.
Application of this strategy to Chromobacterium violaceum has identified a mini-Tn5 luxCDABE Km2 mutant, designated CV113 (Table 1), with an insertion in a gene controlled by HHL (Fig. 4). It is probable, if using traditional methods where the mutants are grown overnight on medium±HHL, that this mutant would be overlooked as one of many dim mutants because final light levels are elevated by only about 1 log by addition of HHL (Fig. 4). The use of Lucy1 has enabled the effect of HHL to be quantified throughout the growth phase, highlighting the induction of gene expression in early exponential phase in the presence of HHL and giving approximately a 2-log difference in reporter bioluminescence (Fig. 4). The use of this type of screen may be of great value where there is no obvious phenotype in the luxI-homologue null mutant, e.g., as in Enterobacter agglomerans.
In experiments designed to study the more general aspects of growth phase dependent gene expression in Aeromonas hydrophila, we identified bioluminescent mutants carrying mini-Tn5 luxCDABE Km2 insertions. A number were selected for analysis using Lucy1 in LB and Minimal M9 broth at 30°C. Growth phase variation in promoter(P)::lux expression from two insertion mutants is shown in Fig. 5. Expression in LB shows fusion has occurred to genes expressed in exponential phase (P18::luxCDABE Km2; Fig. 5A) and in stationary phase (P40::luxCDABE Km2; Fig. 5B). In comparison, the P::lux induction profiles of these mutants grown in Minimal M9 medium shows expression is induced in early exponential phase.
Fig. 6 shows an example of a mutant that would not be as easily characterised using classical methods. In LB, P1::luxCDABE expression is essentially constitutive, with a small induction in stationary phase, whereas in Minimal M9 broth two things are apparent. Firstly, transposon insertion appears to have inactivated a gene required for growth in Minimal M9 and secondly, maximal expression from this P1::luxCDABE Km2 fusion is almost 2 log higher in Minimal M9 than in LB, demonstrating that this apparently essential gene is being induced by the prevailing conditions.
In these experiments it is important to be aware of any effects due to substrate availability. Cells that are starved or stressed may exhibit a lower level of bioluminescence than that reflective of lux gene expression. Conversely, a high level of bioluminescence may cause an energy drain affecting growth rate, bioluminescence signal and cellular viability. A rapid increase in bioluminescence after the addition of fresh substrates (e.g., long chain fatty aldehyde or glucose) is often diagnostic for such problems. The inclusion of a highly bioluminescent, constitutively light clone as a control in these experiments may help identify this type of problem.
The Photorhabdus luminescens luxCDABE containing promoter probe plasmids and mini-Tn5 constructs described in this study provide tools for the construction of reporters of gene expression, which are usable non-destructively in real time and at 37°C. The use of the Lucy1 combined photometer and luminometer in concert with mini-Tn5 luxCDABE mutagenesis now offers the potential to rapidly screen for genes induced at any phase of growth under defined environmental conditions.
This work was funded by grants from Amersham International plc, and the Biotechnology and Biological Sciences Research Council.