Bacterial Bioluminescence: Applications in Food Microbiology

J. M. BAKER', M. W. GRIFFITHS', and D. L. COLLINS-THOMPSON2

Departments of !Food Science and ^'Environmental Biology, University of Guelph, Guelph, Ontario NIG 2W1, Canada Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 (Received for publication June 10, 1991)

ABSTRACT a biochemical and genetic level. The applications of bacte rial bioluminescence will also be presented, with an empha Many marine microorganisms (Vibrio, Photobacterium) are sis on the applications in food microbiology. capable of emitting light, that is, they are bioluminescent. The light-yielding reaction is catalyzed by a luciferase, and it involves BIOCHEMISTRY OF THE the oxidation of reduced riboflavin phosphate and a long-chain BIOLUMINESCENT REACTION aldehyde in the presence of oxygen to produce a blue green light. The genes responsible for the luciferase production, (lux A and lux B), aldehyde synthesis (lux C, D, and E), and regulation of The bioluminescent reaction, catalyzed by the enzyme luminescence (lux I and lux R) have all been identified, and recent luciferase, involves the oxidation of a long-chain aldehyde research has resulted in the discovery of three new genes (lux F, and reduced riboflavin phosphate (FMNH2) and results in G, and H). The ability to genetically engineer dark microorgan the emission of a blue green light. te fc e isms to become light emitting by introducing the lux genes into FMNH2 + 02 + RCOH — ' ™ —> them has opened up a wide range of applications of biolumines FMN + RCOOH + H20 -(-light (490 nm) cence. Assays using bacterial bioluminescence for the detection The primary source of energy for the light is supplied by and enumeration of microorganisms are rapid, sensitive, accurate, the conversion of the aldehyde to the corresponding fatty and can be made specific. It is these attributes that are making in acid. Long-chain aldehydes are essential for the lumines vivo bioluminescent assays so attractive to the food industry. cence reaction and the aldehyde, tetradecanal, appears to be the natural substrate for the luminescence reaction (25). The reaction is highly specific for FMNFL, which is formed by the reaction: Bioluminescent bacteria (those capable of emitting FMN light) are classified into four major genera: Vibrio, NAD(P) H + FMN — °»^^^— > FMNH2 Photobacterium, Alteromonas, and Xenorhabdus. They are Thus, the luciferase reaction may be driven by coupling it classified as motile, Gram-negative rods, and function as to any system that produces FMNH2. Electrons for the facultative anaerobes (25,26). The first three genera are reduction of flavin mononucleotide (FMN) are provided by marine in origin and can be isolated from seawater or the reducing power derived from the electron transport certain luminous ocean species. They can exist free living pathway (75). The mechanism of the reaction leading to in the ocean, as symbionts with various marine fish in their light emission is shown schematically in Fig. 1. FMNH2 digestive tracts or specialized light organs, as saprophytes plus the luciferase results in luciferase-bound FMNH2. on dead fish, and as parasites in Crustacea and insects (26). Luciferase-bound FMNH2 reacts with oxygen to form the The marine organisms coexisting with the luminescent 4a-peroxyflavin intermediate. The intermediate reacts with bacteria can use the luminescence for a variety of purposes a long-chain aldehyde to form an excited species - the such as attraction of prey, communication between species, flavin-4a-hydroxide (72). This compound is highly stable and escape from predators. However, it is not known what and decays slowly resulting in the light emission and the specific benefit these bacteria derive from this energy oxidation of the substrates (26). consuming property, either free living or when associated with higher organisms (77). Detailed biochemical and ge THE ENZYME LUCIFERASE netic studies have been carried out on species of Vibrio and Photobacterium. The results of these studies have led to an Luciferase is the enzyme catalyzing the bioluminescent advancement in the understanding of bacterial biolumines reaction and it is linked to the respiratory pathway. It is an cence. Researchers are now beginning to discover a wide external flavin mono-oxygenase or mixed function range of applications and uses for bioluminescence. This oxygenase. All luciferases are heterodimers (77 kDa) con review will cover the luminescence reaction in bacteria at sisting of two nonidentical subunits a (40 kDa) and [3 (37

JOURNAL OF FOOD PROTECTION, VOL. 55, JANUARY 1992 BACTERIAL LUMINESCENCE 63 hv tern. There are seven genes and two operons required for

NAD(P)H + H A NADPH + HT the complete Lux phenotype. Five of the seven genes are FMN RCOOH "-"•,- ATP structural, coding for the luciferase and the fatty acid reductase. The other two genes are regulatory, responsible FATTY ACID 0X1DO- LUCIFERASE for producing an autoinducer and a receptor. The two REDUCTASE REDUCTASE operons are transcribed in opposite directions. The right | I L AMP+ ppi operon carries genes for the autoinducer, luciferase, and the NAD(P) .J^ FMNH„ RCHO V V NADP"+1 fatty acid reductase. The left operon carries the gene for the receptor protein. Fig. 2 shows the organization of the Vibrio fischeri lux genes.

Figure 1. Schematic representation of the biochemical pathways involved in bacterial bioluminescence. [From Engebrecht el al, 1983 (11)].

Operon Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 kDa) (25,26). The active site is located primarily on the a subunit, but the p subunit is still essential for the light- emitting reaction and may contribute to the structure of the active centre. Neither the a or p subunit alone exhibits Regulation Aldehyde Luciferase Aldehyde -J I I l_ luciferase activity, but both preparations regain activity 0 1 2 3 4 5 6 7 8 9 when combined with the second subunit - leading to the kbp conclusion that an individual subunit does not appear to possess an active centre. Studies indicate that the aldehyde Figure 2. Organization of the Vibrio fischeri lux genes. [Modified from binding site is, like the FMNH2 binding site, at or near the Hastings el al, 1985 (13)]. interface of the luciferase a and p subunits (75). Luciferase genes lux A and lux B ALDEHYDE BIOSYNTHESIS The a subunit is coded for by lux A and lux B codes for the p subunit. Both genes are approximately 1 kbp in Aldehydes are essential in the bioluminescent reaction length and are located immediately adjacent to one another and luminescent cells have a mechanism of synthesizing and are transcribed in the same direction. Aldehyde depen aldehydes from fatty acids. A multienzyme fatty acid re dent luminescence requires only a 2.5 kbp DNA fragment ductase complex that produces aldehydes has been purified with the lux A and B genes under a suitable promoter. Only and characterized from Photobacterium phosphoreum (25). the luciferase genes are necessary for luminescence (25). It The net reaction: is now known that lux A and lux B arose by gene duplica R-COOH + ATP + NADPH tion (6), and there is about 30% identity in amino acid —-> R-CHO + AMP +PPi +NADP sequence between any luciferase a and p subunits regard is the reduction of fatty acid to aldehyde with the oxidation less of bacterial species (4,14,38). of NADPH and the cleavage of ATP to AMP and PPi. Maximum activity is obtained with tetradecanoic acid (25). Fatty acid reductase genes lux C, D, and E Two enzyme activities are necessary for the reaction to The genes coding for aldehyde synthesis, lux C, D, and occur. Acyl-protein synthetase is responsible for the ATP E, must also be transformed and expressed in E. coli to dependent activation of the fatty acid, and a reductase obtain light emission without addition of aldehyde. The lux component catalyzes the NADPH-dependent reduction of C gene encodes for the reductase, the lux D gene encodes the activated fatty acid. Activation of the carboxyl groups for the acyl transferase, and the lux E gene encodes for the is catalyzed by the synthetase and results in an acyl-AMP acyl-protein synthetase (25). intermediate being formed. A NADPH linked reduction of the acyl group to aldehyde is effected by the reductase Regulatory genes lux 7 and lux R (25,26). Acyl-transferase is also present in the fatty acid Two regulatory genes, lux I and lux R, involved in the reductase complex and is involved in the generation of fatty genetic control of luminescence, have been identified in V. acids. Current data suggest that acyl-transferases are in fischeri MJl (11,12). The genes lux F, lux G, and lux H volved in diverting fatty acids into the luminescence system have recently been identified (36,37). Lux F codes for a from the lipid biosynthesis pathway. flavoprotein of about 24 kDa with no known function at this time. Lux G and lux H are both 25 kDa proteins also THE lux GENES of unknown function. Meighen speculates on some possible functions of these new proteins and also provides an up Escherichia colt cells have been transformed with dated series of organizational maps (26). The in-depth cloned genes coding for the luminescence functions and study of luminescent bacteria on a molecular level is expression has resulted in light emission. This has ad starting to result in exciting discoveries about these organ vanced understanding of the gene organization and regula isms and the mechanisms responsible for their unique tion of luminescence in the bacterial bioluminescence sys- characteristic.

JOURNAL OF FOOD PROTECTION, VOL. 55, JANUARY 1992 64 BAKER, GRIFFITHS AND COLLINS-THOMPSON Regulation of luminescence Oxygen is an important and variable factor with respect to A significant portion of the cellular energy of biolumi- the expression of the luminescent phenotype. In some nescent bacteria, (10% or more) can be utilized in produc species such as V. harveyi and P. leiognathi, lux gene ing light (11), and the expression of this phenotype is expression is inhibited by growth at low oxygen. Others tightly regulated by physiological factors. such as P. phosphoreum and V. fischeri exhibit stimulation Autoinduction. In 1977, Nealson (29) reported that the of the luminous system when grown in low oxygen (10). synthesis of the luminous system of Photobacterium fischeri Thus, when using the lux genes from these organisms in (now V. fischeri) was stimulated, at the level of transcrip recombinant studies, the oxygen levels should be monitored tion, by an autoinducer produced by the bacteria them to ensure that changes in luminescence result from the selves. He also found that differences in light intensity were condition or situation under investigation, not fluctuating accounted for by differences in the amounts of autoinducer oxygen levels. As the bacterial luminescent reaction re produced. He noted a lag in luminescence and attributed it quires oxygen, using this method to study anaerobic bacte to two factors. The first factor was a negative one, an ria is not feasible. The presence of iron inhibits luciferase inhibitor that had to be removed. The second factor was a synthesis, and thus light emission, in all species of lumi positive one, the production by the bacteria of an autoinducer nous bacteria studied (10). Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 stimulating the synthesis of the luminous system. It was discovered later that the autoinducer for V. INSTRUMENTATION fischeri is N-p-ketocaproylhomoserine lactone. Analogs of this autoinducer could not stimulate luminescence in other species (25) suggesting some degree of species specificity Light measurements from luminescent reactions (39) in autoinduction. and commercial luminometers have been reviewed (32,41). With the added knowledge of the genes responsible for The sensitivity of commercial luminometers has also been luminescence, there is now a proposed mechanism for assessed (17). autoinduction in V. fischeri involving both positive and Most luminometers rely on sensitive photomultiplier negative feedback regulation. The autoinducer would be tubes for detection and amplification of photon emission. produced initially at a low constitutive rate as a product of The time course of the light emitted by an ongoing lumi the lux I gene on operon R. When enough autoinducer has nescent reaction is the physical property measured. Two accumulated during cellular growth, it would interact with components of light reach the detector. One is proportional a receptor molecule encoded by the lux R gene on operon to the amount of the limiting reactant in the luminescent L and activate transcription of operon R (the lux structural reaction. The other is a fairly constant level of light due to genes). This results in the production of more autoinducer outside causes including reagent impurities and plastic and an exponential increase in synthesis of the lux en phosphorescence. This background "noise" is much lower zymes. High levels of the receptor-inducer complex would than the luminescent component. Some luminometers are then turn off expression of the lux R gene. This explains quite sensitive, being able to detect amounts of ATP as low why luminescence becomes constant or declines at high as 100 femtograms in ATP bioluminescent assays (17). cell density. This model would predict that synthesis of the However, extreme sensitivity may also lead to high stan autoinducer, as well as expression of the structural genes, dard errors when dealing with small cell numbers. would increase during growth (13,25). In situations where In a luminometer, the amount of light produced is the autoinducer cannot accumulate, the bacteria will appear directly related to the total volume in the cuvette. This is very dim as found in organisms living free in the ocean because the instrument reads light equally from all parts of (26). the cuvette. Another key parameter in luminometric mea A recent study (27) showed that the structural genes surement is the rate of reaction of the production of light. (lux ABCDE) of V. fischeri and V. harveyi are highly This rate is directly affected by the concentration of lumi conserved indicating that the light-emitting systems are nescent material in the sample. Diluting a sample slows the very similar in the two bacteria. However, it was also found rate of the luminescent reaction, thus reducing the light that the lux regulatory systems appear to have diverged. In output in photons per second. Measuring the light output V. harveyi there was no open reading frame of greater than (rate of reaction) results in a lower number for the dilute 40 codons within 600 bp of the start of lux C, which is sample. The measurement of light emission is also affected where lux I is located in V. fischeri. This led to the by turbidity and color of the sample. conclusion that either the basic mechanism for the induc Instruments in which photon detection and amplifica tion of luminescence or the location of the regulatory genes tion is achieved by avalanche photodiodes are also com in relation to the structural genes differs in the two Vibrio mercially available. These instruments have a sensitivity systems. approaching photomultiplier-based luminometers but can Other regulatory controls. Catabolite repression par be made much smaller and are more rugged (

JOURNAL OF FOOD PROTECTION, VOL. 55. JANUARY 1992 BACTERIAL LUMINESCENCE 65 CCD has a photosensitive area of 1 cm2. This area contains Transformation and conjugation light-sensitive elements called pixels defined by an array of Transformation involving plasmids to transfer the lu closely spaced electrodes. Photons liberate electrons which minescence genes into a host has also been successful (40). then migrate towards the electrodes within each pixel. The These plasmids contain purified DNA comprising appropri electrons accumulate in the pixels. After a specified time ately modified bioluminescence genes. At high population the charges stored in each pixel may be read out serially by densities (107 to 108 cells per ml) cells of E. coli trans a process called charge coupling. The charge from each formed with specific plasmids exhibited light emission pixel may be used to create an analogue TV picture. within 2 to 5 h (40). However, the time of onset of In an experiment done to assess the potential applica luminescence was inversely proportional to the cell concen tion of CCD imaging in studying luciferase gene expression tration when transformed with a constant amount of DNA. in mammalian cells, a vaccinia virus recombinant contain Conjugation involves the transferring of DNA from one ing firefly luciferase genes was added to monolayers of bacterial cell to another by plasmids containing the transfer CV-1 cells (African Green Monkey kidney cells), and light genes. Transformation and conjugation are the methods detection was measured by the CCD imaging method. After most often used in the transfer of the lux genes into various only 6 h of infection, distinct light-emitting cells could be gram-positive and gram-negative bacteria. Meighen has Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 visualized demonstrating the extreme sensitivity of this summarized the organisms and methods used for the trans method for the early detection of single foci of infection. fer of lux genes (26). Ulitzer and Kuhn (40) found that Data presented indicate that this is a rapid, non invasive, transformation or conjugation of lux DNA was approxi sensitive method of detection which will offer a wide range mately 1000-fold less sensitive for detection of bacteria of potential applications to the study of biological systems than was transduction. (15,16). Early experiments performed on the lux genes con sisted of isolating the genes in the marine organisms and expressing them in E. coli in order to further study the GENETIC TECHNIQUES USED TO luminescent reaction. INVESTIGATE THE lux GENES Belas et al. (7) in 1982 isolated the lux A and lux B genes from V. harveyi and transformed them on plasmids Several researchers (3,7,11,26) working on aspects of pBB123 and pBB128 into E. coli in order to examine the the lux genes have used a variety of genetic techniques in regulatory mechanisms of the genes. They found that light their studies. In order to study luminescence in nonlight- production could be enhanced by coupling the lux genes to emitting organisms, the genes responsible for luminescence strong promoter segments. The lux genes were in fact must be introduced into the dark organism. Transduction, downstream of two bacteriophage promoters. Due to this transformation, and conjugation are all suitable methods of increase in light emission, they felt that the limiting factor introducing DNA (40). for expression of luminescence in E. coli was the availabil ity of substrates - perhaps the reduced flavin. They also Transduction realized the possibility of cloning the genes necessary for Transduction involves the use of phages, which can the aldehyde synthesis in order to eliminate the need to carry the luminescence genes. The phages do not possess supply exogenous aldehyde. the intracellular biochemistry to express the genes but will Miyamoto et al. in a 1987 study (28) transformed the efficiently introduce these cloned genes into an appropriate genes for luciferase and the genes for fatty acid reductase host. Bacteriophages usually have a narrow host range, from V. harveyi into E. coli. Only 6.3 kbp of DNA was growing on some or all strains within a species. By care transformed, which had insufficient coding capacity to fully choosing a narrow host range phage, one can test for synthesize regulatory proteins similar to those found in the the presence of specific bacteria. For example, phages V. fischeri system, thus showing that transcription of regu carrying lux genes whose host ranges cover all known latory proteins is not required for expression of lumines strains of E. coli may be prepared. On infection of the host cence. In past experiments, where V. harveyi DNA contain bacterium, the lux genes are transduced into E. coli which ing the luciferase genes had been cloned into E. coli, only has the necessary intracellular chemistry to express them. very low levels of light were observed. When cells that Detection of this organism is, thus, indicated by the emis expressed high levels of light were found, Miyamoto et al. sion of light. This detection system is rapid and specific (28) investigated to see if expression of the luminescence (40). The difficulty with this methodology is in defining the system arose from an alteration in the recombinant DNA or host range of the phage for different species. Too broad a in the E. coli cells themselves. They found that the E. coli specificity could introduce the genes into species other than had mutated to allow expression of the V. harveyi system. those desired, leading to a false-positive response. Too There was no difference in the recombinant DNA isolated narrow a specificity may not transfer the lux genes into all from the luminescent E. coli and the original recombinant of the strains of the desired species giving a negative DNA. Thus, it was concluded that the host E. coli dramati response (26). Detection of 100-1000 E. coli cells can be cally affects the expression of the luminescence genes of obtained within 1 h in milk or urine (40). Salmonella marine bacteria. species transduced by phages containing the lux A and lux Expressing the bacterial luciferase genes in Bacillus B genes could be detected within 1 hour of infection at cell subtilis as well as E. coli was examined by Karp et al. (20). numbers as low as 10-100 (26). Efficient expression of genes from gram-negative bacteria

JOURNAL OF FOOD PROTECTION, VOL. 55, JANUARY 1992 66 BAKER, GRIFFITHS AND COLLINS-THOMPSON in gram-positive bacilli is not seen as due to more stringent It was found that the fusion of the genes resulted in a good binding of ribosomes and/or to large heterogeneity in RNA level of luciferase production in transformed cells (2). The polymerase sigma factors recognizing different promoter fused gene has successfully been expressed in both yeast regions in these organisms. Solving these expression barri cells and Drosophila melanogaster cells (2). The amount of ers is partially accomplished by fusing foreign genes with light emitted by yeast cells (Saccharomyces cerevisiae) was special vectors which have a strong Bacillus promoter and quite low (1% of the activity in cells without the fusion), Shine-Dalgarno sequence for ribosome binding. and this could be due to the fact that the supply of FMNH2 Karp et al. (20) cloned genes encoding light production is extremely limited in yeast. The researchers felt that the from V. harveyi into a shuttle vector to examine expression amount of light emitted from Drosophila cells should be of lux genes in B. subtilis. They found that light emission sufficient for many kinds of experiments with the organ in B. subtilis was higher when the plasmid containing the ism. Generating in vivo luminescence is more difficult in lux genes had the correct binding site for B. subtilis. They eukaryotes because there is limited availability of FMNH2, also stated that lux genes provided a better way to study although it appears that aldehydes can cross the cell mem expression barriers in B. subtilis because other indicators brane (26). require cell disruption and enzyme assay steps. Further The scope of work done on the lux genes from biolu- Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 work (9,79) showed that B. subtilis containing lux genes minescent bacteria, from identifying the genes and their was found to differ from E. coli in light emission charac products to inserting them into other organisms, all stems teristics. Light emission in E. coli is rapid after aldehyde from a desire to use the emission of light in a wide range addition and stays constant or rises as the cells continue to of applications. The assay is sensitive, quick, relatively grow, reflecting regeneration of substrates for the luciferase inexpensive, does not destroy cell integrity, and can be reaction (19). In B. subtilis, after addition of the long-chain made specific by using host specific vectors. For the above aldehyde, there is immediate light emission - a peak - reasons, coupled with the successes achieved in expressing which, depending on growth phase, slowly decays. The the genes in foreign hosts, a wide array of applications has decay may be due to the obligate aerobic nature of B. been, and still is being developed. subtilis. It appears that it does not have enough capacity to APPLICATIONS OF lux GENE TECHNOLOGY regenerate FMN to its reduced form for the luciferase reaction. However, other reasons may exist for the differ Many different disciplines are investigating the use of ences in light emission between E. coli and B. subtilis (9). lux gene technology. Biomass assays are being developed Two systems were developed by Baldwin et al. (5) to to monitor water quality and for clinical diagnosis of allow quantitative determination of the efficacy of pro bacteruria (31). moter or terminator elements based on light emission. The Work has been done to produce E. coli cells that emit terminator screening vector placed the lux A and lux B stable light thus allowing them to be used as biosensors genes downstream from the promoter with the multiple (22,23). Evidence has been produced that suggests this cloning site between the promoter and the lux genes. new bioluminescent based sensing method could be devel Fragments inserted into the multiple cloning site containing oped for direct analysis of biodegradative microbial activity termination signals caused a reduction in light intensity in in complex environmental matrices such as soil (21). cells carrying this plasmid. The promoter screening vector Ulitzer and Kuhn (40) have used lux genes to deter is similar, but no promoter is present so that cells carrying mine the antibiotic susceptibility of bacteria. E. coli cells the plasmids are dark without an inserted promoter. possessing the lux genes exhibited decreased light emission Two papers were published in the early 1980s (11,12) in the presence of certain antibiotics. Suppression of lumi dealing with the identification of the genes and their prod nescence depended upon the cell's susceptibility to and the ucts responsible for bacterial luminescence. Various ge concentration of the antibiotic. This is of great importance netic techniques were used in order to define and character in diagnostic medical microbiology. A commercially avail ize the genetic information. Determination of the genetic able toxicity test, Microtox, based on bacterial biolumines- organization and definition of the lux functions was per cence has the potential of replacing animal testing. formed by transposon mutagenesis. Transposons mutate by Lux genes have been inserted into the broad host range insertional inactivation of the target gene causing loss of vector pUCD4 resulting in the recombinant plasmid function. Nonluminous and dim mutants were isolated and pUCD607. The recombinant plasmid was mobilized into transposon insertions were located by restriction mapping. plant pathogenic bacteria. This allowed bacterial invasion The regions encoding aldehyde, luciferase, and regulatory in host plant tissues to be quantified by light emission even functions lux genes A to E and lux I and lux R were before the onset of visual symptoms (30). This may have identified. These functions were assigned to the genes by potential for monitoring storage life of fruits and veg examining the phenotypes of mutants. etables. These are just a few examples of the many appli Very recent work (2) with lux genes has involved the cations now being found for bacterial bioluminescence. fusion of the lux A and lux B genes into a lux AB gene on the plasmid pHG 165-3M. This fusion is necessary for APPLICATIONS OF lux GENE TECHNOLOGY facilitated insertion of the genes into eukaryotic cells, since IN FOOD MICROBIOLOGY expression in eukaryotes requires separate promoters before each gene. This fusion is expected to lead to an increase in In a review of the potentials of in vivo biolumines applications for the luciferase genes in eukaryotic systems. cence in microbiology, Stewart states three components he

JOURNAL OF FOOD PROTECT/ON, VOL. 55. JANUARY 1992 BACTERIAL LUMINESCENCE 67 feels are necessary for novel developments in lux gene typhimurium could be detected in 50 min. by monitoring technology (34). The first is the potential of a high level of light expression (35). Lux containing bacteriophage should light output from individual cells. The second is the rapidly be able to target other pathogens such as Campylobacter expanding ability to transfer the biochemistry of light spp. and Listeria monocytogenes. There may still be a need production by the cloning of lux gene vectors into normally for recovery and enrichment procedures prior to phage dark microorganisms. The third component involves the detection, because of regulatory requirements, for example use of instruments originally designed for in vitro ATP zero tolerance for L. monocytogenes. However, due to the assays to allow the detection of light from only a few sensitivity (100 per ml) and the specificity of the host and hundred bacteria per ml. All of these are important in phage, the enrichment time could be short providing simple implementing bacterial bioluminescent assays into the realm same day testing for pathogenic bacteria (34). of microbiological testing. Whereas a number of ATP High numbers of nonpathogenic microorganisms present luminescent techniques are presently being employed in within a food can indicate an increased probability of various aspects of food microbiology, the application of pathogen contamination. Recombinant lux+ phages can de bacterial bioluminescence is still on the research or proto tect these indicator bacteria without recovery or enrichment type level. Many experiments have been, and are being

as long as they are in a food matrix at levels greater than Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 done, to determine the feasibility and usefulness of the 1000 per g. The assay takes only 30-50 min, thus allowing proposed assays. Some of the applications of use to the for an index of food safety in less than 1 h (34). food industry include the detection of specific bacterial Enteric bacteria are used in the food industry as indi pathogens and indicator microorganisms, monitoring hy cators of poor sanitary conditions. Lux genes have been giene on-line, determining the effectiveness of spore de inserted into phages that infect a broad range of enteric struction, monitoring starter culture integrity, biocide and bacteria providing a reagent for an on-line hygiene test with virucide challenges, and studying recovery of sublethally a detection limit of 1000 enterics per g (35). This test injured cells. would take less than an hour to perform, would not require capital equipment, and could be operated by factory per Detection of specific bacterial pathogens and indicator sonnel. This would allow an accurate assessment of the organisms numbers of enterics in raw materials, the product during Presently in food microbiology laboratories, the mea manufacture, swabs from the floor and equipment, and the surement of ATP by firefly luciferase is used to detect and end product. The test has shown promise in prototype form enumerate cells and has been used to assess shelf life and (18,35). It would be particularly useful in the monitoring of the microbial quality of many kinds of foods (24,33). The critical control points as part of the Hazard Analysis Criti ATP assay is a rapid technique, but it lacks specificity for cal Control Points (HACCP) system, whereby defined criti identification of bacteria present. This specificity can be cal control points in a manufacturing process are to be achieved with lux genes from luminescent bacteria. As rapidly monitored by physical, chemical and microbiologi discussed previously, lux genes can be introduced into cal indicators. Presently, microbiological monitoring results bacteriophages which will then absorb to specific bacteria require from 8-24 h. As many hazards are microbiological, and transfer the light-emitting genes to those bacteria. By it would be very advantageous to rapidly sample the micro knowing what type of bacteria one wishes to detect, it is bial flora. If these bacteriophages can be engineered to only a matter of obtaining the host specific phages for that provide a reagent capable of detecting changing levels of particular organism, performing the genetic manipulations, organisms in an hour then perhaps on-line microbial testing combining the organisms with the phages, and measuring will become part of the HACCP system (35). the light emission, to obtain a quantitative assay of the organism. Research has shown that the bioluminescent Sporeforming organisms and bioluminescence assay is rapid (usually less than 1 h), sensitive [luminometers can detect as few as 500 bacteria (35)], simple, specific, Bacterial luciferase will function in a gram-positive and demonstrates a good correlation (linear relationship) organism although it is originally from gram-negative or between cell numbers and bioluminescence. ganisms. The lux genes must be coupled to appropriate gram-positive promoters and must be engineered to obtain The dark terrestrial organisms that need to be moni stable insertion via gram-positive replicons (regions con tored in food microbiology (pathogens, starter cultures, taining DNA necessary for the initiation of DNA replica hygiene indicators) lack the ability to produce luciferase or tion) (35). Bacillus spp. are capable of producing heat fatty acid reductase, but they can supply FMNH . There 2 stable dormant endospores. The spores obtained from phe- fore, all that is needed is the transfer of the genes for notypically bioluminescent vegetative cells are dark. This is luciferase and fatty acid reductase (35). In some cases only expected because the spores show no detectable metabo the genes for luciferase are transferred and the aldehyde is lism, ATP or electron transport, thus having no energy to supplied exogenously as it can cross the cellular membrane. drive the light reaction. When the spore germinates, the By this means, one can confer bioluminescence on a par onset of electron transport and the initiation of metabolism ticular set of bacteria growing in a complex mixture of are early events. For lux containing spores, germination is microbial organisms. Phage P22 is a narrow host range accompanied by the emergence of bioluminescence. This phage infecting only Salmonella typhimurium. Stewart and provides a sensitive, real-time monitor of the germination co-workers have constructed P22 containing lux genes. On and outgrowth process (35). infection with this phage, as few as 100 cells of S.

JOURNAL OF FOOD PROTECTION, VOL. 55, JANUARY 1992 68 BAKER, GRIFFITHS AND COLLINS-THOMPSON Spores that have been killed or injured, and are thus correlation with the viable count evaluation (18,34). Since unable to germinate, produce no light. Therefore, those dead cells produce no light, the biocidal effect is measured chemical and physical processes designed to eliminate as a decrease in the light output of the bioluminescent bacterial spores could be monitored within 1 h in a culture. These results may lead to rapid on-site evaluation luminometer assay. Researchers feel that a rapid biological of biocide efficacy in process water control and water- indicator for thermal processing or sterilizer efficacy moni holding tanks where bacteria are known to proliferate toring could be available in the near future (35). (34,35). This method has been used to test the effectiveness A similar approach for studying germination of spores of certain biocides on L. monocytogenes with good results of Clostridium spp. would be advantageous. However, after 15 min exposure to the biocides (3). The genes for problems may be encountered due to the anaerobic nature luminescence have been introduced into L. monocytogenes of the organism. cells by transforming them with a recombinant plasmid. The plasmid (pSB292) is a broad host range gram-positive Lactic acid bacteria and starter cultures vector which replicates in L. monocytogenes and thus A plasmid based replicon pSB 154 has been engineered provides a means for incorporation of the lux genes into the to introduce bioluminescence into a wide range of lactic Listeria cells. It was demonstrated that bioluminescence Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 acid bacteria including Lactococcus lactis, Lactobacillus closely correlates to cell growth of L. moncytogenes until casei, and Lactobacillus plantarum. Only the lux A and lux the approach of stationary phase when bioluminescence B genes are expressed, so aldehyde must be supplied for begins to decay (3). bioluminescence (35). Levels of light emission from these Assessing antiviral activity usually requires cell culture organisms are less than those from the enteric bacteria work or electron microscopy facilities (18). A rapid biolu recombinants, but it is possible to monitor in real-time minescent test has been developed by inserting lux A and concentrations of 106 bacteria per ml. Further research is lux B genes into Lambda phage. After exposure to hypo expected to result in improvements in quantum output, but chlorite (virucidal agent), the lux* phages were unable to useful reagents have been developed. Many of these assays infect a nonbioluminescent E. coli and so did not elicit a have a relation to the dairy industry. The presence of bioluminescent response. Untreated lux* phages did pro bacteriophages or antibiotics can cause starter culture fail duce a bioluminescent phenotype in E. coli even after the ure in cheese or yogurt manufacture. These agents impair E. coli had been treated with hypochlorite. This demon growth of the starter and allow contaminating microorgan strates the potential of using lux* phages to screen for isms to proliferate. Dye reduction tests are used currently to virucidal activity. evaluate antimicrobial activity in milk, but these require several hours of incubation before a result is obtained. Recovery of sublethally injured cells Bioluminescent lactic streptococci are suited for use as Sublethal injury refers to the state where a microorgan indicators of the presence of lytic phages or antibiotics. ism is neither actively growing or biochemically dead. This light-emitting component of the starter culture can be Microorganisms in this state may arise in a food after monitored for a loss of luminescence that indicates the heating, chilling, freezing, moisture reduction, irradiation, presence of an inhibitory substance (26,34,35). Using a or exposure to preservatives. Sublethally injured cells can bioluminescent derivative of Lactobacillus casei, the tech retain their pathogenic traits which makes enumeration of nique allowed detection of penicillin G down to levels of these cells during microbial analysis very important. They 0.03 |Xg/ml (0.05 units/ml) in 30 min and bacteriophages at usually need a period of resuscitation in a nonselective concentrations as low as 105/ml in 100 min (1). medium for intracellular repair (18). For Salmonella, an Certain species of lactic acid bacteria are currently overnight resuscitation in an appropriate preenrichment being investigated as probiotics in the control of intestinal medium is needed before any attempt to determine its flora in poultry and pigs. Bioluminescent bacteria could presence is made (35). This compromises efforts to reduce play a role in assessing the colonization and attachment to the time for bacterial detection and enumeration based on the gut epithelium of organisms such as Lactobacillus new rapid technologies. Protocols used currently to monitor ingested live at substantial levels (35). recovery are based on observing colonies obtained under different resuscitation conditions and overnight incubation Effectiveness of biocides and virucides (18). Researchers now believe that in vivo bioluminescence Biocides are used for microbiological control in manu may provide a real-time tool for probing the recovery of facturing, environmental, engineering, service, and other microorganisms from sublethal injury. This injury affects industries as preservatives or disinfectants (J8). Material the quantitative efficiency of in vivo bioluminescence. It is such as food waste, organic soil, dust and other organisms possible with sensitive luminometers to watch the recovery can neutralize biocide activity without reducing apparent of sublethally injured cells to optimum bioluminescence chemical activity (35). For this reason, monitoring of bio without having to observe a single cell division. Work done cide levels is essential to provide adequate protection. indicates that three hours into the recovery period the Current testing techniques (dipslides, dye reduction tests, viable count data and bioluminescence reflect near perfect microbiological challenge tests) require in excess of 18 h correlation (18). Different protocols for recovery can be and are thus not suitable for on-site evaluation. Biolumines assessed with great sensitivity both to optimize chemical cent microorganisms offer the potential for rapid biocide conditions and to reduce the time required (35): soon it testing. The test takes only 10-15 min and gives a good should be possible to investigate the effect of media,

JOURNAL OF FOOD PROTECTION, VOL. 55, JANUARY 1992 BACTERIAL LUMINESCENCE 69 temperature, and other environmental parameters on recov correlation with cell counts - both when increasing and ery in vivo (18). Recovery from sublethal stress is an decreasing in numbers. When using narrow range phages to emerging concern in food microbiology. Any method to introduce the lux genes, the assay can be specific to detect advance its understanding should be pursued with vigor. a particular microorganism. A final advantage is that the Stewart (34) deals with how in vivo bioluminescence technology of bacterial bioluminescence is leading to a can be a reporter of microbial stress by acting as a reporter great deal of innovation in many areas. New ideas are of gene expression. The luciferase expression is a very emerging that were not thought possible before, such as versatile reporter gene. It does not require any special actually watching the recovery of an injured cell. genotype in the host strain in order to obtain expression As with any new technique, the researchers developing data and the response is obtained in real time without cell the technology are slow to mention too many disadvan disruption. This is important when exploring cellular re tages. As mentioned in Meighen's review (26), there could sponse to environmental stimuli. Another application inves be a problem with phage or plasmid host ranges being tigates the capacity of cells to supply high energy bio either too specific or too broad, resulting in false negatives chemical intermediates (i.e., FMNH2) during or after some or false positives, respectively. More research into phages type of cellular stress. Stress responses that affect the and plasmids and their host specificity should rectify this Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 production of intracellular FMNH2 can be monitored in lux problem. Bioluminescent cells have been detected in milk, recombinant bacteria as changes in light output per cell but one can envisage a setback with solid foods, in that any (34). For example, using bioluminescent recombinant strains particulate matter may decrease the efficiency of gene of S. typhimurium, workers have shown that 20% of the transfer within the food homogenate. Sample pretreatment population could survive a freeze/thaw cycle with a bio such as filtration may be a solution to this problem. How chemical system sufficiently intact to permit immediate and ever, further research needs to be done in this area. sustained bioluminescence (34). By classical plate tech A disadvantage for workers in the food microbiology niques it was found that only 2.5% of the original inoculum industry is that the bacterial bioluminescent assays are still remained viable. This has important implications for the at the experimental and prototype stages. food industry as these results suggest that appreciable numbers of toxigenic cells (e.g., S. aureus) that are not SUMMARY considered viable by conventional enumeration techniques may continue to produce toxigenic compounds after pro A great deal of research has been done in the last 10 cessing. to 15 years to investigate the mechanisms whereby marine The monitoring of bioluminescent cells exposed to bacteria bioluminesce. From an initial understanding of different stresses may find application in the field of pre their biochemistry and the substrates required for light dictive microbiology. Trying to predict how an organism is emission, to investigation of their molecular biology and going to behave in a certain environment is not an easy defining what genes code for what enzymes, to genetically task, though many are pursuing it. By inflicting a variety of engineering otherwise dark microorganisms to become light stresses on different components of the cell's metabolism emitting, these organisms have become very well defined. and then monitoring light expression, one can see what Along with the basic scientific investigation of these organ conditions are harmful or beneficial to the organisms. isms, many applications of their luminescence are being realized, thus leading to further research on these unique ADVANTAGES AND DISADVANTAGES microorganisms. The many advantages of assays using OF THE BACTERIAL LUNINESCENT SYSTEM bacterial luciferase should increase their use, especially in IN FOOD MICROBIOLOGY food microbiology, once they become commercially avail able. These assays will not eliminate classical microbiology The advantages of assays performed with bacterial methods, but they will further the influence of microbial luminescent genes have been repeated throughout this re assays into areas such as the HACCP system where rapid view. The assays are rapid — fast enough for on-line results are necessary. Advancements in genetically engi assays in situations where on-line monitoring was impos neering microorganisms to become bioluminescent coupled sible previously, and also fast enough for assays that with a desire for rapid microbiological assays by the food require immediate action, such as the detection of antimi industry will make in vivo bioluminescence a common crobials in milk. The rapidity may also in some cases practice in the future. Bioluminescence coupled with low reduce costs as well. For example in testing the efficacy of level light detection systems such as photon counting and virucides, there would be no need for expensive tissue CCD imaging (techniques capable of visualizing biolumi culture supplies. nescent single cells on food surfaces) will provide the food The method is simple enough to operate or use so that microbiologist with a very powerful tool for studying food plant workers could perform the test. It is a sensitive as an ecosystem and will offer a unique opportunity to method - very low numbers of cells can be detected with study microbial behavior and interactions in food. luminometers and continued research may lower those numbers even further. In most cases it is advantageous that REFERENCES the in vivo bioluminescence assay is noninvasive. No cell 1. Ahmad, K. A., and G. S. A. B. Stewart. 1991. The production of disruption is required to measure the light emission. It has bioluminescent lactic acid bacteria suitable for the rapid assessment been shown to be an accurate method exhibiting good of starter culture activity in milk. J. Appl. Bacteriol. 70:113-120.

JOURNAL OF FOOD PROTECTION, VOL. 55, JANUARY 1992 70 BAKER, GRIFFITHS AND COLLINS-THOMPSON 2. Almanshu, S., B. Musafia, R. Hadar, M. Suissa, and J. Kuhn. 1990. subtilis. pp. 361-364. In J. Scholmerich, R. Andreesen, A. Kapp, M. Fusion of Lux A and Lux B and its expression in E. coll, S. Ernst, and W. G. Woods (ed.), Bioluminescence and chemilumines cerevisiae, and D. melanogaster. J. Biolum. Chemilum. 5:89-97. cence new perspectives. John Wiley and Sons, Toronto. 3. Anonymous. 1990. The introduction of a bioluminescent phenotype 21. King, J. M. H., P. M. DiGrazia, B. Applegate, R. Burlage, J into Listeria monocytogenes. AFRC News. Sanseverino, P. Dunbar, F. Larimer, and G. S. Sayler. 1990. Rapid, 4. Baldwin, T. O., L. H. Chen, L. J. Chlumsky, J. H. Devine, T. C. sensitive bioluminescent reporter technology for naphthalene expo Johnston, J. W. Lin, J. Sugihara, J. J. Waddle, and M. M. Zeigler. sure and biodegradation. Science 249:778-781. 1987. Structural analysis of bacterial luciferase. pp. 621-631. In E.E. 22. Korpela, M., and M. Karp. 1988. Stable-light producing Escherichia Ednondson and D. B. McCormick (ed.), Flavins and flavoproteins. coli. Biotech. Lett. 10:383-388. Walter de Gruyter and Co., Berlin. 23. Korpela, M., P. Mantsala, E. Lilius, and M. Karp. 1989. Stable-light- 5. Baldwin, T. O., J. H. Devine, J. W. Lin, R. Legocki, A. Szalay, D. emitting Escherichia coli as a biosensor. J. Biolum. Chemilum. S. Peabody, and D. G. Bear. 1987. Applications of the cloned 4:551-554. bacterial luciferase genes lux A and lux B to the study of transcrip 24. Kricka, L. J. 1988. Clinical and biochemical applications of lucif- tional promoters and terminators, pp. 373-376. In J. Scholmerich, R. erases and luciferins. Anal. Biochem. 175:14-21. Andreesen, A. Kapp, M. Ernst, and W. G. Wood (eds.), Biolumines- 25. Meighen, E. A. 1988. Enzymes and genes from the lux operons of cence and chemiluminescence new perspectives. John Wiley and bioluminescent bacteria. Ann. Rev. Microbiol. 42:151-176. Sons, Toronto. 26. Meighen, E. A. 1991. Molecular biology of bacterial biolumines 6. Baldwin, T. O., M. M. Ziegler, and D. A. Powers. 1979. Covalent cence. Microbiol. Rev. 55:123-142.

structure of subunits of bacterial luciferase: NH2-terminal sequence 27. Miyamoto, C, M. Boylan, L. Cragg, and E. Meighen. 1989. Com Downloaded from http://meridian.allenpress.com/jfp/article-pdf/55/1/62/2303097/0362-028x-55_1_62.pdf by guest on 30 September 2021 demonstrates subunit homology. Proc. Natl. Acad Sci. USA. 76:4887- parison of the lux Systems in Vibrio harveyi and Vibrio fischeri. J. 4889. Biolum. Chemilum. 3:193-199. 7. Belas, R., A. Mileham, D. Cohn, M. Hilmen, M. Simon, and M. 28. Miyamoto, C, D. Byers, M. Boylan, A. Graham, and E. Meighen. Silverman. 1982. Bacterial bioluminesccnce: Isolation and expres 1987. Vector and cell-dependent expression in Escherichia coli of sion of the luciferase genes from Vibrio harveyi. Science 218:791- the genes responsible for bacterial bioluminescence. pp. 389-392. In 793. J. Scholmerich, R. Andreesen, A. Kapp, M. Ernst and W.G. Wood 8. Brown, R. G. W. 1989. ATP photon counting using avalanche (eds.), Bioluminescence and chemiluminescence new perspectives. photodiodes. pp. 257-259. In P. E . Stanley, B. J. McCarthy, and R. John Wiley and Sons, Toronto. Smither (ed.), ATP luminescence rapid methods in microbiology. 29. Nealson, K. H. 1977. Autoinduction of bacterial luciferase. Arch. Blackwell Scientific Publications, Ltd., Oxford. Microbiol. 112:73-79. 9. Carmi, O. A., G. S. A. B. Stewart, S. Ulitzer and J. Kuhn. 1987. Use 30. Shaw, J. J., and C. 1. Kado. 1986. Development of a Vibrio biolumi of bacterial luciferase to establish a promoter probe vehicle capable nescence gene-set to monitor phytopathogenic bacteria during the of nondestructive real-time analysis of gene expression in Bacillus ongoing disease process in a non-disruptive manner. Biotechnology spp. J. Bacteriol. 169:2165-2170. 4:560-564. 10. Delong, E. F., D. Steinhauer, A. Israel, and K.H. Nealson. 1987. 31. Stanley, P. E. 1981. Overview of applications of bioluminescence. Isolation of the lux genes from Photobacterium leiognathi and pp. 275-280. In M. A. Deluca and W. D. Mc Elroy (ed.), Biolumi expression in Escherichia coli. Gene 54:203-210. nescence and chemiluminescence. Basic chemistry and analytical 11. Engebrecht, J., K. Nealson, and M. Silverman. 1983. Bacterial applications. Academic Press, New York. bioluminescencc: Isolation and genetic analysis of functions from 32. Stanley, P. E. 1982. Instrumentation, pp. 219-260. In L. J. Kricka Vibrio fischeri. Cell 32:773-781. and T. J. N. Carter (ed.), Clinical and biochemical luminescence. 12. Engebrecht, J., and M. Silverman. 1984. Identification of genes and Marcel Dekker, New York. gene products necessary for bacterial bioluminescence. Proc. Natl. 33. Stannard, C. J., and P. A. Gibbs. 1986. Rapid microbiology, appli Acad. Sci. USA. 81:4154-4158. cations of bioluminescence in the food industry- a review. J. Biolum. 13. Hastings, J. W., C. J. Poitrikus, S. C. Gupta, M. Kurfurst, and J. C. Chemilum. 1:3-10. Makemson. 1985. Biochemistry and physiology of bioluminescent 34. Stewart, G. S. A. B. 1990. In vivo bioluminescence: New potentials bacteria, pp. 236-291. In A. H. Rose and D. W. Tempest (ed.), for microbiology. Lett. Appl. Microbiol. 10:1-8. Advances in microbial physiology, vol. 26. Academic Press, Toronto. 35. Stewart, G., T. Smith, and S. Denyer. 1988. Genetic engineering for 14. Haygood, M. G. 1990. Relationship of the luminous bacterial sym- bioluminescent bacteria. Food Sci. Techno]. Today 3:19-22. biont of the Caribbean flashlight fish, Kryptophanaron alfredi (fam 36. Swartzman, E., S. Kapoor, A. F. Graham, and E. A. Meighen. 1990. ily Anomalopidac), to other luminous bacteria based on bacterial A new Vibrio fischeri lux gene precedes a bidirectional termination luciferase (lux A) genes. Arch. Microbiol. 154: 496-503. site for the lux operon. J. Bacteriol. 172:6797-6802. 15. Hooper, C. E., and R. E. Ansorgc. 1990. Quantitative luminescence 37. Swartzman, E., C. Miyamoto, A. Graham, and E. Meighen. 1990. imaging in the biosciences using the CCD camera: Analysis of Delineation of the transcriptional boundaries of the lux operon of macro and micro samples. Trends Anal. Chem. 9:269-277. Vibrio harveyi demonstrates the presence of two new lux genes. J. 16. Hooper, C. E., R. E. Ansorge, H. M. Browne, and P. Tomkins. 1990. Biol. Chem. 265:3513-3517. CCD imaging of luciferase gene expression in single mammalian cells. J. Biolum. Chemilum. 5:123-130. 38. Szittner, R., and E. Meighen. 1990. Nucleotide sequence, expression 17. Jago, P. H., W. J. Simpson, S. P. Denyer, A. W. Evans, M. W. and properties of luciferase coded by lux genes from a terrestrial Griffiths, J. R. M. Hammond, T. P. Ingram, R. F. Lacey, N. W. bacterium. J. Biol. Chem. 265:16581-16587. Macey, B. J. McCarthy, T. T. Salusbury, P. S. Senior, S. Sidorowicz, 39. Turner, G. K. 1985. Measurement of light from chemical or bio R. Smither, G. Stanfield, and P. E. Stanley. 1989. An evaluation of chemical reactions, pp. 43-78. In K. Van Dyke (ed.), Biolumines the performance of ten commercial luminometcrs. J. Biolum. cence and chemiluminescence: Instruments and applications, vol. I. Chemilum. 3:131-145. CRC Press, Inc., Boca Raton, FL. 18. Jassim, S. A. A., A. Ellison, S. P. Denyer and G. S. A. B. Stewart. 40. Ulitzer, S., and J. Kuhn. 1987. Introduction of lux genes into 1990. In vivo bioluminesccnce: A cellular reporter for research and bacteria, a new approach for specific determination of bacteria and industry. J. Biolum. Chemilum. 5:115-122. their antibiotic susceptibility, pp. 463-472. In J. Scholmerich, R. 19. Karp, M. 1989. Expression of bacterial luciferase genes from Vibrio Andreesen, A. Kapp, M. Ernst, and W. G. Wood (ed.), Biolumines harveyi in Bacillus subtilis and in Escherichia coli. Biochim. Biophys. cence and chemiluminescence new perspectives. John Wiley and Acta 1007:84-90. Sons, Toronto. 20. Karp, M. T., P. Meyer, M. Lahdc, R. Lahti, and P. I. Mantsala. 1987. 41. Van Dyke, K. 1985. Commercial instruments, pp. 83-128. In K. Van Construction of a shuttle plasmid containing bacterial luciferase Dyke (ed.), Bioluminescence and chemiluminescence: Instruments genes for gene expression studies in Escherichia coli and Bacillus and applications, vol. I. CRC Press, Inc., Boca Raton, FL.

JOURNAL OF FOOD PROTECTION. VOL. 55, JANUARY 1992