FIELD OF THE INVENTION The present invention relates to an assay to determine the level of sensitivity a DN detection method can achieve, such a DNA detection method being one useful in the detecting th presence of bacteria that carry a particular DNA sequence in a large heterogenous population o bacteria which do not contain the particular DNA sequence.

BACKGROUND OF THE INVENΗON The introduction of transcriptionally active recombinant DNA into the environment, eithe intentionally or during the manufacture of recombinant proteins, is a safety concern of both th biotechnology industry and government regulatory agencies including the Food and Dru Administration and the Environmental Protection Agency. Capture of transcriptionally active DN by naturally occurring microorganisms could result in replicating bacteria different from th originally engineered strain and capable of producing the recombinant protein which the capture DNA encodes. Bacteria can transfer or exchange genetic material by mating, by infection b bacterial viruses called bacteriophages, or through other means of taking up genetic material fro the environment. There is concern that genetic material could eventually be transferred to bacteri which are indigenous to animal enteric tracts whereby such recipient bacteria could survive an produce proteins encoded by the newly acquired genetic material. In addition, there is a concer that engineered strains of bacteria could colonize the enteric systems of animals directly. In order to determine the liklihood that genetic material will be transferred from a engineered strain to endogenous microorganisms associated with an animal or that a non endogenous engineered microorganism could survive in an animal, a great deal of effort has been expended studying the exchange of genetic material between bacteria and the survival in an animal's enteric system of bacteria that contain transferred DNA. These studies involve exposing an animal to bacteria containing a gene of interest and then detecting the presence of the gene o interest in the microbial flora of an organism sometime after the time of exposure. According to these studies, the detection of the gene of interest indicates that either the engineered bacteria has colonized the subject animal or that the gene of interest originally possessed by the engineered bacteria is transferred to a bacterium capable of surviving in the animal; such a bacterium most likely being an endogenous bacteria.

The methodology used in these studies generally involve oral exposure to the engineered bacteria followed by screening the bacteria in the fecal material of the animal for bacteria containing the gene of interest. Oral administration corresponds to the likely route bacteria enter the animals body. Bacteria regularly inhabit the intestinal tracts of animals and genetic transfer between engineered and endogenous species or colonization by engineered strains will likely take place there. The examination of microbial flora in fecal matter is a sound method of surveying the

inhabitants of an animal's gut.

One shortcoming for detecting the presence of a gene of interest in bacteria that exist in a heterogenous population is that the sensitivity of such assays is either poorly defined or not defined at all. Accordingly, the ill-defined limits of sensitivity make interpretation of the results of these experiments unreliable. It is an arduous task to validate the assay to ensure that the observations scored either positive or negative are precise and accurate. Moreover, the assays currently employed lack a well defined range of sensitivity in which one can precisely and accurately define the sensitivity of the signal chosen for detection.

According to the present invention, the process of gene detection in heterologous populations can be improved upon by choosing a suitably sensitive detection method and by providing a means to validate the method's limits of sensitivity. One such means of approach to the binary problem of sensitivity and verification is to generate an organism to serve as a model the organism which may result from the transfer of heterologous DNA. Such a model organism could serve as an important assay standard and could be conveniently and expeditiously engineered by the use of known transducing bacteria phages. The use of such an organism coupled with a sensitive gene detection technique, e.g., DNA hybridization, could be used as a validated assay system. Accordingly, data generated by such assays will provide a guide to the accuracy and range of sensitivity of detection assays used to determine the presence and/or movement of genes of interest in bacteria which inhabit animal hosts. The data generated in the detection assay is more reliable and useful when sensitivity limits are established, providing the reviewer of such data an improved opportunity to draw conclusions necessary for making informed decisions.

IMORMATTON DISCLOSURE Cohen, P.S., etal., Recomb. DNA Tech, Bull., 2:106-113 (1979), refer to the colonization potentials of male and female E. coli K12 strains E. coli B and human fecal E. coli strains in the mouse GI tract. Enteric flora is eliminated in mice by providing them with drinking water containing streptomycin. The mice are then fed various E. coli strains that contain multiple selectable markers and the survival of the strains is determined by culturing bacteria from the mouse fecal matter under selective pressure.

Levine, M.M., et al., Journal Infectious Diseases 148:699-709 (1983) refer to the recombinant DNA risk assessment studies in humans; the efficacy of poorly mobilizable plasmids in biologic containment. Bacterial strains are used that contain selectable markers which render them sensitive under certain selectable conditions and which also contain plasmids with selectable markers that render them resistant under certain selectable conditions. In this experimental model, the bacteria are orally introduced in human subjects. Bacteria from the stools of the subjects are cultured under conditions that select against the introduced bacteria strains but select for bacterial population carrying the introduced plasmids. Bacteria which grow under such conditions are

scored as endogenous bacteria which received the resistance selectable plasmid by transconjugatio Stotzky, G. and H. Babich, Recomb. DNA Tech. Bull., 7:163-188 (1984), refer to the fat of genetically-engineered microbes in natural environments. Studies involving the survival o recombinant microbes and DNA are reviewed. The studies relate to determining the presence o recombinant bacteria that contain selectable markers and the presence of DNA cont∑dning selectabl markers.

Smith, C, Jr., et .., Recomb. DNA Tech. Bull., 8:47-51 (1985), refer to the effect o colonizing mice with laboratory and wild type strains of E. coli containing tumor virus genomes Laboratory and wild type strains of E. coli which contained recombinant polyoma virus DNA ar introduced into mice. The mice are evaluated for the development of polyoma infection. Th presence of the recombinant bacteria is determined by plating enteric bacteria derived from th mice selecting for selectable markers present in the recombinant strains. Additionally hybridization assays are performed detecting the presence of the polyoma DNA. The sensitivit of the hybridization assay is neither disclosed or discussed. George, S.E., et al., Environmental Toxicology and Chemistry 1:123-131 (1989), refer to the acute colonization study of polychlorinated biphenyl -degrading pseudomonades in the mous intestinal tract; a comparison of single and multiple exposures is made. Recombinant bacteria strains administered to mice are detected in enterobacteria isolated after sacrifice by culturing unde selective conditions. SUMMARY OF THE INVENTION

The present invention relates to the use of lysogen comprising a single copy of a gene o interest and a selectable marker to determine the sensitivity of an assay, detect a gene of interes in a heterogenous population of bacteria. In addition, the present invention relates to a method o determining the level of sensitivity of an assay that is used to detect bacteria carrying a gene o interest in a heterologous population of bacteria. The method comprises the steps of producing a mixed population of bacteria by mixing a population of lysogens that comprise a gene of interest and a selectable marker with a population of non-lysogens having neither the gene of interest nor the selectable marker. The number of lysogens and the number of non-lysogens in the mixed population is determined and the mixed population is plated under conditions which select for presence of the selectable marker present. The plated colonies are additionally checked for the presence of the gene of interest.

DETAILED DESCRIPΗON OF THE INVENTION To practice the present invention, a lysogen is constructed that contains a gene of interest and a selectable marker. The lysogen can be generated from a strain of bacteria capable of surviving in the gut of an animal which is to be the subject of in vivo studies. However, the choice of strains is not restricted if the sensitivity assay is to be performed solely as an in vitro assay.

As used herein, the term "lysogen" refers to a bacterial strain in which a foreign piece of

DNA is inserted into the chromosomal DNA of the bacteria in a specific location. Lysogens are generated by infecting a bacterial cell with a recombinant bacteriophage (phage) that contains a gene of interest and a selectable marker. The phage DNA integrates into the bacteria's chromosome.

As used herein, the term "gene of interest" refers to a gene which is to be tracked or detected. Often, a gene of interest encodes a desired protein and the gene of interest is inserted in a production host in order to produce the protein by recombinant means.

Generally, assays are performed to determine the capability of a gene of interest's to move from one host strain to another, the latter strain being one capable of surviving in an animal host. Alternatively, assays are performed to determine whether a bacterial strain which carries a gene of interest is capable of surviving in a host animal. In either case, a gene of interest, usually within the production host, is introduced into a potential animal host, usually orally. The microbial flora in the animal's fecal matter is then examined for the presence of the gene of interest. If the gene of interest is located, either it has been taken up by an endogenous bacterial strain or the introduced strain is capable of surviving in the host animal's gut.

The assay used to detect the gene of interest comprises growing the fecal matter-derived bacteria and employing means to indicate the presence of the gene of interest. Various detection means are available. The most common and straight forward way to detect the presence of a gene of interest is by hybridization using a radiolabelled probe. A probe comprises a radiolabelled copy of the gene of interest or a portion thereof. Production of probes useful in hybridization techniques are well known to those having ordinary skill in the art.

According to this method, the fecal matter-derived bacteria are grown in single colonies using agar media. A single colony is a clonal colony where every cell contains identical genetic material and each cell in the colony is derived from a single parent cell. As used herein, the term "plating" refers to the techniques for growing single colonies. Plating techniques are well known to those having ordinary skill in the art. Briefly, bacteria is diluted into a range of dilutions and added to petri dishes or plates which contain agar media. At the proper dilution the bacteria are diluted sufficiently such that they are separated from each other at a g at enough distance so that resulting colonies are discrete and all cells in a colony are derived from the single cell. Techniques involving plating on nitrocellulose paper and performing colony lifts onto nitrocellulose paper are also well known to those having ordinary skill in the art.

In order to determine the level of sensitivity that detection assays are capable of attaining, the assay according to the present invention is developed. The in vitro hybridization assay described here is developed to address in vitro the impact of plasmid transfer in a heterogenous population of mouse gut E. coli.

The present invention relates to the use of a lysogen to determine the sensitivity of an assa which is used to detect a gene of interest in a heterogenous population of bacteria. The presen invention also relates to a method of determining the level of sensitivity of an assay which is use to detect bacteria carrying a gene of interest in a heterologous population of bacteria. Th techniques used to practice the present invention are well known to those having ordinary skill i the art and can be performed using readily available starting materials.

To practice the present invention, a lysogen is constructed which contains a gene of interes and at least one selectable marker. It is, however, preferable that the lysogen contain tw selectable markers, one of which is closely adjacent to the gene of interest. The lysogen i constructed by infecting a bacterial strain with a bacteriophage which does not enter the lyti growth cycle, i.e., the presence of the phage in the host does not result in the production o infective particles. Rather, the phage DNA integrates into the bacterial chromosomal DNA an is referred to as a prophage. It is preferred that the bacteriophage carries both a copy of the gen of interest and a selectable marker. If the gene of interest and the selectable marker are adjacent genetic recombination of the lysogen can be determined. Genetic recombination of the phage DN in such lysogens is indicated if the lysogens are detected using antibiotic selection but do no contain the gene of interest as detected by a hybridization assay. If a second selectable marker i provided, it can be present anywhere in the bacteria cell, either on the chromosome or on plasmid within the host cell. The presence of two selectable markers allows for the selection o a small number of slow replicating lysogens when mixed with a larger number of non-lysoge bacteria.

A lysogen starter culture and non-lysogen starter culture are grown up. Various known quantities of lysogen starter culture are diluted and mixed with various known quantities of non- lysogen starter culture. The mixtures of cultures, referred to herein as the "mixed populations", are plated on nitrocellulose filters over agar plates that contain selection media useful to select for lysogens. If the lysogen contains two selection markers, the agar plates contain media that selects for both markers.

In order to determine how many lysogens and how many non-lysogens are mixed together to form the mixed population, the starter cultures are both plated on non-selection agar and colonies are grown up and counted. Determining the concentrating bacteria in culture by plating is routine. Once the concentration is determined, the number of lysogens mixed with non-lysogens when the two are mixed can be determined.

The colonies which grow on the selection plates from the plating of the mixed population are lysogen. The colonies are counted. Finally, nitrocellulose lifts of the mixed population plates are made and tested by hybridization assay using a probe to the gene of interest.

Since the number of lysogens and non-lysogens can be calculated for every mixed

population plate, the data generated from the colony counts and hybridization assay can be used to determine the level of sensitivity of the detection assays. That is, comparing the expected results which can be calculated with the actual results which are observed, the level at which these two sets of data agree is the level of sensitivity of the detection assay. Beyond the level of sensitivity, observed results cannot be considered reliable. However, within the level of sensitivity, results can be accepted as accurate and true.

Attaining this level of sensitivity in mixed cultures is feasible only by using antibiotic selection to favor lysogen growth to a colony size which would allow sufficient quantities of denatured DNA to be fixed to the nitrocellulose. By using antibiotic selection, the present invention detects the presence of a gene of interest at a level of 1 colony of bacteria that contain a single copy of a gene of interest per bacteria out of 1 X 10 ⁸ background colonies. EXAMPLE

The capture of transcriptionally active DNA from recombinant production strains of bacteria by naturally occurring microorganisms and the replication of such wild type bacteria thus transformed and capable of producing the recombinant protein for which the captured DNA encodes is a concern raised when assessing the safety of recombinant DNA technology in the production of commercially valuable peptides. The possibility that such an event can occur is of interest because of the development of recombinant bovine somatotropin (rBSt) as a commercial product. In order to determine the likelihood of this event as well as the persistence of this phenomena occurring in nature, a validated murine gastrointestinal tract model colonized with E. coli carrying transcriptionally active BSt cDNA is used.

The experimentation designed to address these environmental concerns necessitated the use of a bacterial strain capable of colonizing the murine gastrointestinal tract and modified to carry the BSt cDNA in a single chromosomal copy. The use of a lysogenized strain which carried both the antibiotic resistance of the non-lysogenized parent strain as well as an alternative antibiotic resistance to the parent strain conferred by the prophage has the added benefit of selecting for the alternative antibiotic marker and/or the BSt insert. The added advantage of the double antibiotic resistance of the lysogen is the ability to select for the lysogen out of a background population of the colonized parent strain. Additionally, the prophage antibiotic resistance marker could be used to determine if the BSt insert have segregated from the host chromosome. Thus, a primary screen for antibiotic resistance could be established followed by the secondary hybridization screen.

In initial in vivo experimentation using bacterial colonies isolated from murine fecal material of mice inoculated with streptomycin resistant non-lysogenized and streptomycin/ampicillin resistant lysogen strains showed a detection limit of 1 positive in a background of approximately 5 X 10 ⁴ non-lysogens. While this limit of detection may be sufficient for the in vivo assay as defined by the number of colonies which could be obtained from a 100 mg sample of mouse fecal

material, it is felt that the large fermentation volumes generated at the production level togethe with the anticipated low frequency of this event occurring in nature necessitated the establishmen of a lower limit of detection that could be addressed statistically through an in vitro design. Th in vitro experimentation described here has two important advantages over the in vivo assay in tha the size of the bacterial population as well as the ratio of the lysogen to nonlysogen strains coul be controlled carefully.

The ability to detect 1 positive out of a population of 5 x 10 ⁵ plaque forming units i feasible when conducting hybridization lifts during lambda gtll cDNA library screening. However, that library screen is conducted on 22 cm ² plates containing 2 x 10 ⁵ pfu as compare to the 100 mm diameter plates containing 5 x 10 ⁴ cfu used during the in vivo animal mode hybridizations. Colonies seeded at this density grow as well isolated clones and are able to gro to an average colony size of 0.5 mm which is sufficient for detecting a strong hybridization signal. However, due to colony size limitations imposed by the 100mm agar plates when seeded at 1 10 ⁶ cfu/plate and the presence of BSt cDNA at only one copy per cell, the amount of BSt cDN per colony bound to the nitrocellulose membrane is extremely low, thus minimizing or eliminating detection. The in vitro experimentation according to the present invention is designed to assess th sensitivity of the hybridization assay and whether or not this level of sensitivity could b determined from lysogens grown on both non-selective and selective media.

The embodiment of the present invention described herein consists of three components; 1) use of an existing bacterial phage lambda to create a phage containing a copy of the bovine somatotropin (BSt) gene; 2) use of the bovine somatotropin gene carrying phage to infect wildtype E. coli cells creating BSt phage lysogens; and, 3) use of the BSt lysogens in an existing murine animal model to demonstrate colonization of the GI tract by the cells and subsequent use of these strains in this model as a component of an environmental assessment study required by the Federal Food and Drug Administration for the approval of bovine somatotropin.

Materials used in one embodiment of the present are described below. Reagents used in the preparation of solutions are of reagent grade or better. Ampicillin and streptomycin as well as Kodak XAR-5 film for autoradiography is purchased from Sigma Chemicals, St. Louis, Mo. MacConkey agar is purchased from Becton Dickinson, Cockeysville, MD, and prepared as per manufacturers recommendations. Falcon 100 X 15 mm polystyrene petri dishes are purchased from Becton Dickinson Labware, Lincoln Park, NJ. Nitrocellulose filter membranes (S&S #20460 BA85; .45um pore size; 82.5mm diameter) for DNA blotting is obtained from Schleicher and Schuell, Keene, NH. Random primer labelling reagents are obtained from Stratagene, La Jolla, CA. ³² P-dCTP is obtained from Amersham, Chicago, IL. Whatman 3MM chromatography'paper is purchased from VWR Scientific, Piscataway, NJ. Analysis of variance using a 2 X 2 X 2 factorial design is conducted using SAS software, copyright 1984,1986; SAS Institute Inc., Cary,

N.C. MacConkey agar plates are prepared as per manufacturers instructions. In some cases the molten media is supplemented with either streptomycin (100 /ig/ml) or streptomycin (100 μg/ml) and ampicillin (100 μg/ml). The entire contents of all references cited below are hereby incorporated by reference. Construction of lysogens is described below. First, a phage carrying the BST cDNA is constructed. It is then used to generate a bacteria lysogen.

An existing λ phage, BDC531 is obtained from Dr. Donald Court, Frederick Cancer Institute, Frederick MD. This phage is employed in a homologous recombination experiment with cells transformed with the BSt plasmid, pLEBGH-Z. The BSt cDNA is described in Tomich, C-S. C, et al., Plasmid 20:167-170 (1988). From this experiment, a new population of phage carrying the bovine somatotropin gene are created. Purified BSt phage are subsequently modified in a fashion which prevented their entering a lytic growth cycle when present as prophage in the bacterial chromosome. Phage generated in this series of experiments are assigned the laboratory designations XBST21 and XBST434. Infective λ phages BST21 and BST434 are used to generate lysogens in E. coli cells isolated from murine gastrointestinal tract. The E. coli isolates employed are designated and are . made resistant to the antibiotic streptomycin by selection of spontaneous mutants on streptomycin plates. Using standard phage methodology, three E. coli isolates are infected in separate experiments with XBST21 and XBST434 phage to generate a total of six lysogenic strains capable of colonizing murine gastrointestinal tract. Cells lysogenized with XBST434 are preferred for in vivo applications because of their genotype which prevents prophage from entering a lytic growth cycle.

The ability of the hybridization assay to detect XBST434 lysogens designated strains UC12692 and UC12693 in a background of non-lysogens designated strains UC12699 and UC12700 is assessed by plating mixed populations of lysogens and non-lysogens on nitrocellulose filters over on agar plates containing selective media and subsequently probing filter lifts prepared from these plates with radiolabelled BSt cDNA. The inocula for these mixtures are prepared by inoculating beads from frozen stocks into brain heart in&sion broth (BHIB) containing either 100 μg streptomycin/ml (strains UC12699 and UC12700) or into BHIB containing 100 μg streptomycin ml and 10 μg ampicillin/ml (strains UC12692 and UC12693). These cultures are incubated at 36° C without agitation for 16-18 hours before dilution and mixing. The exact inoculum size is determined for each culture by viable plate counts on plain MacConkey agar medium plates.

Non-lysogenized strains UC12699 and UC12700 are diluted to 2.9 and 3.6 X 10 ⁸ cfu/ml, respectively, and mixed with a population of the corresponding XBST434 lysogen (UC12692 or

UC12693) diluted 1.3 X 10 ^"6 , 1.3 X 10 ^"7 , and 1.3 X 10 ^"8 . The plates for the hybridization assays

are prepared by inoculating 0.1 ml of the mixtures directly onto 82.5 mm nitrocellulose membrane filters which have been overlayed on MacConkey agar supplemented with both ampicillin (100 μg/ ml) and streptomycin (100 μg/ml). Ten plates per dilution are prepared and these are incubated 16-18 hours at 36° C. Original inoculum densities for each culture is determined by dilution and plating onto plain MacConkey agar. Colony counts are recorded directly from these same plates which are also used for the hybridization assay. The bacterial lawn grew on the surface of th nitrocellulose filter itself thereby eliminating much of the smearing associated with performing a filter lift from cells grown on an agar surface. Ten plates for each lysogen dilution are prepared in this fashion on MacConkey ampicillin/streptomycin agar. These plates are designated Master Plates from which duplicated filters are prepared and probed with the radiolabelled BSt insert. Duplicated filters are prepared from the Master Plates and grown at 37° C for 2 hours in the presence of streptomycin and ampicillin.

Following overnight growth at 36°C, viable colony counts are obtained by direct visual inspection of 0.5 - 1.0 mm diameter colonies. After obtaining the colony counts, one nitrocellulose lift is conducted on each of the ampicillin/streptomycin plates containing the nitrocellulose membranes and grown at 37° C for 2 hours on ampicillin/streptomycin MacConkey plates. The nitrocellulose membranes on which the bacterial mixes have been grown are used as one of the duplicate membranes for the hybridization assay so a direct comparison between colony counts and hybridization signal could be made. Membranes containing the bacterial colony lifts are lysed, neutralized, and washed according to the method of Sambrook, J., et al., "Molecular Cloning: A Laboratory Manual, Coldspring Harbor Laboratory Press, Cold Spring Harbor, N.Y. (1989) by successively placing filters for 5 minutes onto 3MM paper saturated with: 1) 0.5N sodium hydroxide/1.5M sodium chloride; 2) 0.5M Tris-hydrogen chlorideφH 7.4 at 25°C)/1.5 M sodium chloride; 3) 2X SSC. Following the SSC wash, the filters are dried at room temperature for 1 hour, sandwiched between sheets of 3MM paper, baked for 1.5 hours at 80° C/25 inches of vacuum, and stored in the dark at room temperature until screened for the presence of BSt cDNA. Screening of all colonies for BSt cDNA employed the Clal/BamHI restriction fragment of plasmid pURA-4 (pUC1195) containing the BSt coding sequence as described in European Patent Publication 0 418 219. This fragment is radiolabelled to a specific activity of 1-2 X 10 ⁹ cpms/μg using the random primer method of Feinberg, A. P and Vogelstein, B., Anal. Biochemistry, 132:6-13 (1983) as modified for use in the Prime-It Kit™ (Stratagene, LaJolla,CA.). Treatment of the baked nitrocellulose filters prior to and during hybridization is done according to the method of J. Sambrook et al., (1989). Briefly, the baked filters are floated on the surface of 2X SSC until thoroughly wet from beneath and then submerged for 5 minutes. The filters are transferred to a glass crystallizing dish containing 500 mis of 5X SSC/0.5% SDS/lmM EDTA (pH 8.0) and prewashed for 30 minutes

at 50°C. For this and all subsequent steps the filters are agitated at 65 RPM in a New Brunswick G-24 rotating platform incubator. In order to reduce background hybridization following the 30 minute prewash, bacterial debris are gently scraped from the surface of the filters using a camel hair paint brush soaked in prewashing solution. The filters are then transferred to a glass crystallizing dish and prehybridized for 1.5 hours at 68°C in prehybridization solution containing 6XSSC and 0.05X BLOTTO (Johnson, D. A., et al., Gene Anal. Tech., 1:3-8 (1984)). Prehybridization volume is set at 0.2 mis for each square centimeter of filter surface or 10.6 mis per 82.5 mm filter. The ³² P random primer labelled Clal/BamHI restriction fragment is heat denatured at 100°C for 5 minutes followed by rapid chilling in ice water. The heat denatured probe is added to the prehybridization solution covering the filters at a concentration of 1 X 10 ⁶ cpms/ml and rotated at 68°C for 20 hours. Following hybridization the filters are washed successively in 2 liters (4 X 500 ml washes for 5 minutes/wash) of 2X SSC/0.1% SDS at 22°C, 500 mis of IX SSC/0.1% SDS for 1 hour at 68° C, and 500mls of 0.2X SSC/0.1% SDS for 1 hour at 68° C. Following the final wash, twenty filters are grouped according to treatment on 14 x 17 inch sheets of 3MM paper, dried at room temperature, and exposed to x-ray film against a Cronex™ Lightning Plus intensifying screen (DuPont, Wilmington, DE) at -80° C. The autoradiographs are developed and the number of positive colonies determined by direct visual counting.

Lysogens from the bacterial mixtures plated directly on nitrocellulose filters overlayed on ampicillin/streptomycin MacConkey media grew as well isolated colonies. As determined by direct visual inspection, the hybridization counts and viable colony counts for the experiment are listed in Table I. Control plates showed a strong hybridization signal for both the UC12692 and UC12693 lysogens while the UC12699 and UC12700 non-lysogen strains did not exhibit a hybridization signal. Using the selective pressures of combined ampicillin streptomycin selection described herein, the sensitivity of detection is determined to be 1 in 3.6 X 10 ⁸ . Attaining this level of sensitivity in mixed cultures is feasible only by using antibiotic selection to favor lysogen growth to a colony size which would allow sufficient quantities of denatured DNA to be fixed to the nitrocellulose.

The results show that plating the bacterial mixtures directly onto nitrocellulose filters to selectively favor lysogen growth during the initial incubation on ampicillin/streptomycin MacConkey media allowed detection of well isolated lysogen colonies. The ability to conduct hybridization assays on the same colonies from which the viable colony counts are originally obtained, rather than repetitiously performing multiple lifts from the original plate, offers not only the advantage of expedition by reduced handling of the original inoculation, but more importantly reduced the statistical variance associated with the other methodologies used in which separate plates are prepared for visual counting and for hybridization. The experiment disclosed here further defined the sensitivity of the hybridization assay as being able to detect one XBST434 lysogen

within a non-lysogen background population of approximately 4 X 10 ⁸ cfu. This is importan because it extends the operational range of this type of assay to a level well below that require for in vivo experiments designed to address the question of potential plasmid transfer in heterogenous population of E. coli.

Table I - Viable Colony Counts and Hybridization Signals.