Transposon mutagenesis of Serratia marcescens: molecular biology concepts for the undergraduate teaching laboratory.

The red-pigmented bacterium Serratia marcescens has been used as a model organism for a variety of new microbiology and genetic laboratory exercises. In this work we expand the utility of S. marcescens Nima to include more modern approaches of in vivo molecular biology. We describe a method for transposon-mediated reporter gene mutagenesis which creates transcriptional fusions of S. marcescens genes to a bacterial luciferase operon. The resulting mutants express light in proportion to cell number during exponential growth. Phenotypic screening of twenty mutants revealed 45% luminescence-positive fusions. Growth and luminescence profiles of several mutants illustrate the value and complexity of mutant phenotype analysis. This mutagenesis exercise demonstrates a wide variety of basic molecular biological concepts. Equipment required for this exercise includes a waterbath-shaker and a spectrophotometer (both of which are standard for most biological laboratories) and an inexpensive luminometer.

Introduction

As repositories of heritable biochemical information, genes are foundational to the study of total organism biology. Investigations involving gene mutation, regulation of expression, structure and function reveal valuable information that contributes to a detailed understanding of phenotypic characteristics. Many such investigations begin with random mutagenesis to identify genes of interest by a loss of function phenotype. Transposable element biology has been exploited for this purpose for the last quarter century, and modern genetic constructions function simply enough to be useful for advanced undergraduate laboratory exercises. In this work we introduce transposable elements and their research applications to the advanced undergraduate microbiology or genetics classroom.

Transposable elements or transposons are a major source of genome modification and transfer of genes to other organisms (2). Transposons are segments of genetic material, usually DNA, which catalyze their own insertion at new places in the host cell genome either by excision and movement or by replication into a second site. This feature of transposable element biology has been exploited by bacterial geneticists as a means of creating random loss-of-function (null) mutations. Recombinant DNA technology has allowed experimenters to substitute genes of known function for nonessential transposable element sequences. This has made possible the in vivo fusion of genes encoding easily-assayable products to target genes of unknown function. In addition to the elimination of target gene function, the fusion of transposon DNA physically marks the genetic locus and, importantly for this work, provides a quantitative measure of target gene expression. Genes encoding assayable products fused to other genes therefore became known as "reporter genes" because, under certain circumstances, assayable product expression is proportional to activity of the target gene. Transposon-mediated reporter gene mutagenesis has been used extensively in the field of bacterial pathogenesis to identify loci upregulated by mammalian body temperature and other environmental cues such as calcium concentration (4).

Further, transposon-mediated fusions to genes which catalyze light production are used in environmental microbiology to create biological sensors of toxic chemicals such as benzene (1).

Transposons which create genetic fusions are of two basic types (Fig. 1). Transcriptional or operon fusions carry reporter genes without their native transcriptional initiation sites. These promoterless elements rely on the transcriptional start site of the disrupted gene to express the reporter. Therefore, reporter expression can reflect the transcriptional activity of the target gene under a given set of environmental conditions.

[FIGURE 1 OMITTED]

Reporter genes provide an accurate measure of target gene expression provided two assumptions hold true. First, the experimenter assumes that the disrupted gene does not regulate its own expression. This assumption is not always valid, particularly in bacterial systems. Since gene expression can be regulated at transcriptional, posttranscriptional, translational and posttranslational levels, results can sometimes be misleading. In spite of these limitations, reporter gene technology does allow the experimenter to genetically mark a locus with a unique, recoverable DNA sequence. This is often a necessary first step toward gene cloning and more refined regulatory studies in nonmutant cells and functional analysis by means of site-directed mutagenesis.

Serratia marcescens strain Nima expresses the red, intracellular pigment prodigiosin (5). This normally nonpathogenic organism has been useful for the development of undergraduate laboratory exercises demonstrating mutation to antibiotic resistance (6), the regulation of gene expression (7), and classical genetic studies based on analysis of mutant pigmentation phenotypes (8). This work extends the utility of S. marcescens by describing a simple method for transposon-mediated reporter gene fusion mutagenesis. This method may be adapted to the undergraduate microbiology or genetics curriculum as a laboratory aid to teaching the concepts of molecular genetics.

Materials And Methods

Bacteria, media and culture conditions. Serratia marcescens Nima ATCC #29632 was obtained from Carolina Biological Supply Company. Escherichia coli SM10 lambda pir (pUTminiTn5Kmlux) (3) was provided by Dr. John Coote. The control strain E. coli HB101 has been described previously (8). E. coli HB101(F' Tn3) was constructed by conjugal transfer of F' lac from E. coli EG312 (Escherichia coli Genetic Stock Center strain #6217) with selection for 25 [micro]g/mL ampicillin resistance ([Ap.sup.r]). Bacteria were stored in 40% glycerol as 20[degrees]C frozen stocks. Bacteria were cultivated in trypticase soy broth (TSB; Difco Inc., Detroit, MI) supplemented with 1.5% agar as needed for growth on solid medium. LB agar was purchased from Midwest Scientific Company (St. Louis, MO). The S. marcescens pigmentation medium nutrient broth (NB) + 0.5% maltose has been described elsewhere (7). S. marcescens Nima and its derivatives were incubated at 30[degrees]C, and the E. coli strains were incubated at 30[degrees]C to 37[degrees]C. Bacterial cell concentration in broth culture was measured as the optical density at 620 nm (O[D.sub.620]) using a Beckman DU 640B spectrophotometer. Bacterial strains for the gene fusion mutagenesis procedure described below are available from Dr. Pryce L. Haddix upon request.

Mutagenesis procedure. Plate or slant growth of E. coli SM10 lambda pir (pUTminiTn5Kmlux) and S. marcescens Nima was transferred into separate 10 mL TSB tubes and resuspended to slight turbidity. 150 [micro]L of E. coli and 50 [micro]L S. marcescens were added to the surface of LB agar plates and mixed together by spreading over the surface to dryness. The plates were incubated about 24 hours at 30[degrees]C. Inoculum mixing on distinct plates ensured independent solid-phase matings. After lawn growth had appeared on the nonselective plates, approximately 1/3 of the growth was scraped from the surface using a sterile, disposable L-shaped spreader (Midwest Scientific). The growth was thinly spread onto the surface of LB agar plates with 40 [micro]g/mL Km and 25 [micro]g/mL Sm and allowed to incubate for two more days at 30[degrees]C. Growth from the selection plates was picked for single-colony isolation on the same selective medium. Individual isolates were designated MTN5-X, where X is a number. Incubated plates may be held in the refrigerator at any time during or after this procedure for up to one month.

Tests of transposon vector carriage by mutant Serratia strains. Two indirect tests were employed to ensure that each [Km.sup.r] S. marcescens strain did not harbor the transposon delivery vehicle pUTminiTn5Kmlux. Both are tests for expression of the beta-lactamase (B1a) protein, which confers [Ap.sup.r] on cells expressing it. Characterization of the [Ap.sup.r] phenotype was performed by growing the cells to stationary phase in TSB at 30[degrees]C and diluting them in cold PBS on ice (6) for recovery of single colonies on LB agar and LB agar + 25 [micro]g/mL Ap. The second method was direct assay of Bla activity from stationary phase cells grown in TSB at 30[degrees]C (8).

Measurement of bacterial bioluminescence. Bacterial bioluminescence mediated by luxCDABE gene products produces green light at 490 nm (9) and is constant per cell during exponential growth (10). Luminescence of 1 mL culture aliquots was cumulatively measured as relative light units (RLU) over a 90 second read time using an Inspector luminometer (UltraSource, Inc., Peachtree City, GA). In all cases light emission was measured in TSB medium to minimize cellular pigmentation and its potential impact through light absorbance. All measurements were corrected for background luminescence contributed by the medium, which was generally less than 1000 RLU and insignificant at high cellular optical densities.

Statistical definitions and methods. Luminescence activity units are defined as 1000 multiplied by the slope of the regression line from a plot of log luminescence vs. early log phase O[D.sub.620]. Correlation squared values for these plots exceeded 0.9000. Another descriptive parameter, measured peak luminescence, is defined as the highest ratio of luminescence to optical density recorded during log or stationary phase. Since luminescence could not be monitored continuously throughout the growth cycle, measured peak luminescence values are approximate.

Least-squares linear regression, analysis of doubling time data for normality and graph construction were performed using Axum 6.0 statistical software (Mathsoft, Inc.). Broth culture doubling times were calculated from the graphs as previously described (7).

Results

Mutagenesis of S. marcescens Nima with miniTn5Kmlux. The luxCDABE operon of the terrestrial bacterium Photorhabdus luminescens is carried on the mobilizable, [Ap.sup.r] plasmid pUTminiTn5Kmlux. These promoterless genes lie near one end of a modified Tn5 transposon which also encodes kanamycin resistance ([Km.sup.r]). Transcriptional fusions delivered by inter-species conjugation were recovered by selection for [Km.sup.r] and the [Sm.sup.r] unique to S. marcescens. Mating controls, consisting of conjugal donor and recipient plated separately on selective medium containing Km and Sm, produced no growth (data not shown). We conclude that spontaneous mutation of strain Nima to [Km.sup.r] does not occur under the conditions of these experiments and that proof of transposon carriage by molecular methods (e.g., Southern blotting or the less informative PCR technique) is not required for this classroom exercise.

We attempted to simplify the above procedure by mixing liquid cultures of the parental strains and allowing them to incubate at 30[degrees]C without agitation for periods up to two days. Unfortunately, we were not able to recover [Km.sup.r] S. marcescens mutants using this approach (data not shown). Other investigators using similar mutagenesis systems have been able to perform liquid-phase matings using cells centrifuged to form pellets and incubated overnight (B. M. Applegate, personal communication). Because we intended to keep requirements for instrumentation to a minimum, we did not test this procedure.

Loss of transposon delivery vector through its failure to replicate in the S. marcescens host was confirmed by ampicillin sensitivity and negative [beta]-lactamase activity in two representative isolates (Table 2). MTN5-4 and MTN5-5 were the only mutants to appear nonpigmented on LB agar.

Luminescence and growth phenotypes of independent [Km.sup.r] Nima mutants. Twenty colony-purified isolates from twenty separate conjugation platings were screened for luminescence. Even though assignment of Lux phenotype was sometimes complicated by endogenous luminescence from strain Nima, clear phenotypes could be determined for more than half of the strains. Mutants with relatively unchanged luminescence activity and measured peak luminescence were considered non-fusions. Mutants with increases in both of these parameters were considered positive lux fusions. A potentially confounding scenario involving mutational elimination of a negative regulator of background luminescence was considered unlikely. Those mutants with decreases in one or both parameters were designated unknown.

As Table 3 indicates, the mutants displayed a wide variety of phenotypes.

Table 3 shows a 1.6-fold range in luminescence activity and a 4.4-fold range in peak luminescence among the 9 mutants designated Lux+. Mutants MTN58 through MTN5-10 showed strong consistency with the Nima luminescence and doubling time parameters. In fact, n = 3 t tests for these three parameters revealed no significant differences between Nima (three identical measurements) and its negative fusion mutants (one measurement each of three different strains).

A normal plot of the mutants' doubling time data revealed an overall negligible effect of mutation on cellular physiology (Figure 2). Sixteen of the 20 mutants showed a normal distribution about their mean of 36.0 minutes with a 95% confidence interval of 34.8 minutes to 37.1 minutes. The n = 3 mean of the parental strain Nima is 33.9 minutes. Taken together, these data show a statistically significant but small increase in mutant doubling time by approximately 6%. Pairwise comparisons of the three variables in Table 1 showed a weak positive correlation between mutant doubling time and measured peak luminescence (r = 0.74). The strains identified by number in Figure 2 appear to be defective for nonessential genes which influence general features of Serratia biology.

[FIGURE 2 OMITTED]

Luminescence profiles of selected mutants. Growth and luminescence phenotypes for three mutant strains grown at 30[degrees]C in TSB are presented in Figures 3 through 5. We will assume that the hybrid mRNA molecules encoding luminescence gene products have the same stability in all mutant genetic backgrounds. If this assumption is valid, luminescence measurements directly reflect reporter gene transcriptional activity.

[FIGURE 3 OMITTED]

Figures 3 through 5 show the general trend of exponentially increasing luminescence during log phase. This corresponds to the exponential increase in cell number and suggests that luminescence per cell is constant throughout log phase. Indeed, this observation is the basis for calculation of luminescence activity for comparison of transcriptional activity among strains (see Materials and Methods). Figure 3 shows endogenous luminescence activity of a nonfusion mutant. Activity during both logarithmic and stationary phases was dependent on cellular multiplication. Figure 4 records luminescence of strain MTN5-4 with indeterminate fusion activity. In this case luminescence continued to increase logarithmically through recorded stationary phase. The relatively small rate of luminescence increase reflects a lower luminescence activity than the nonfusion strain MTN5-9 (Figure 3 and Table 1). In addition, two large spikes of luminescence activity are visible and correspond to changes in growth rate. Finally, the positive fusion strain MTN5-15 (Figure 5) showed high luminescence activity and peak luminescence.

[FIGURES 4-5 OMITTED]

Discussion

Our screen of twenty independently-selected luxCDABE transposon mutants identified 45% Lux positive, 15% Lux negative, and 40% of indeterminate fusion phenotype. Of the 12 mutants whose phenotypes could be assigned, 9 or 75% are positive and 3 or 25% are negative. These data fit perfectly a binomial probability model which assumes 50% positive fusions (on the basis of transcriptional orientation) and two transposition events per isolate: 25% of all mutants have two positive fusions; 50% have one positive and one negative fusion; and 25% have two negative fusions. Selection schemes using minimal medium (7) to select against mutations in metabolic genes would probably face a drastic reduction in recoverable colonies per selection plate.

Experience with other genera of Gram-negative bacteria, Pseudomonas and Spirillum, has shown that similar luxCDABE mutagenesis protocols are possible in these organisms (E L Haddix, unpublished data). Pseudomonas fluorescens mutants express background luminescence similar to Nima, while Spirillum mutants express about tenfold higher luminescence per cell with an inducible luxCDABE fusion (1) when compared to Pseudomonas. Unfortunately, Spirillum doubles about every two hours in NB medium at 30[degrees]C, making it a poor choice for classroom exercises.

Other experiments may be performed with the MTN5-X mutants created in this exercise. Growth in minimal medium and growth at 37[degrees]C downregulate Nima pigmentation, while growth in LB and nutrient broth with maltose upregulate pigmentation (7). Light production among the mutants will likely vary in response to these growth conditions, since luminescence is controlled by different cellular operons in the different isolates. More advanced independent study projects are also possible. Investigations into the physiological role of prodigiosin pigment, which absorbs light at 499 nm (7), may be possible with controllable light production at the nearby wavelength of 490 nm.

We present here a simple method for transposon-mediated reporter gene fusion mutagenesis which may be performed over several undergraduate laboratory periods with convenient holding points. This system requires a one-time investment in a single instrument, a luminometer (11), but does not require reagents to assay fusion gene activity. Theoretical topics which are illustrated by this laboratory exercise include: conjugation, transposition, bacterial genetic structure and transcriptional organization, and the biological importance of repeated DNA sequences by means of their in vivo rearrangements. In summary, this exercise helps to illustrate many concepts which are foundational to modern molecular biology and genetic engineering.

TABLE 1
SUGGESTED TASK OUTLINE FOR UNDERGRADUATE LABORATORY APPLICATIONS *

       Session                                Approximate Time Require
Week   Number    Task                         for Class of Ten Students

1      1         Bacteria conjugation on      0.5 hour
                 nonselective media
       2         Selection of K[m.sup.r]      0.5 hour
                 S[m.sup.r] exonjugants
2      3         Colony purification of       0.25 hour
                 independent mutants
       4         Growth and luminescence      4 hours minimum with time
                 curves of Nima (control)     points measured every 0.5
                 and MTN5-X mutants           hour
3      5         Discussion of experimental   1 hour
                 results

* Assumes twice weekly laboratory sessions.

TABLE 2
TESTS OF MTN5-1 AND MTN5-4 FOR VECTOR CARRIAGE

Strain                 Mean CFU on LB Agar    Mean CF U on
                                              LB Agar + Ap

                       [bar.x]                [bar.x]
                       [+ or -] s (n)         [+ or -] s (n)

E. coli H101           ND                     ND
E. coli H101 (F'Tn3)   96 [+ or -] 13 (17)    11 [+ or -] 5 (20)
S. marcescens Nima     374 [+ or -] 31 (10)   0 [+ or -] 0 (10)
S. marcescens MTN5-1   397 [+ or -] 36 (10)   0 [+ or -] 0 (10)
S. marcescens MTN5-4   337 [+ or -] 41 (10)   0 [+ or -] 0 (10)

Strain                 Beta-lactamase
                       Activity, Units

E. coli H101           12.3
E. coli H101 (F'Tn3)   121
S. marcescens Nima     19.0 *
S. marcescens MTN5-1   10.5
S. marcescens MTN5-4   12.1

ND = not done; * Data from reference 8.

TABLE 3
PHENOTYPIC PROPERTIES OF mini-Tn5Kmlux S. marcescens NIMA MUTANTS

MTN5-X Strain       Luminescence   Measured Peak
                    Activity       Luminescence,
                                   RLU/O[D.sub.620]

1                   1355           46345
2                   1070           8792
4                   763            18692
5                   1674           21045
6                   1488           34201
7                   1941           77606
8                   1019           19091
9                   1093           22845
10                  1047           21377
11                  1231           78265
12                  1417           56616
13                  1296           122859
14                  1595           59690
15                  1410           151798
16                  756            31724
17                  606            31487
18                  531            26263
19                  649            29757
20                  1362           106028
21                  848            37926

S. marcescens       1145           23109
Nima
(n = 3)

E. coli             1490           2577
SM10 [lambda] pir

MTN5-X Strain       Early Log      Lux
                    Phase          Fusion
                    Doubling       Phenotype
                    Time/minutes

1                   33.4           +
2                   35.1           ?
4                   34.1           ?
5                   38.9           ?
6                   39.7           +
7                   39.0           +
8                   36.9           -
9                   33.4           -
10                  38.5           -
11                  35.3           +
12                  36.2           +
13                  33.7           +
14                  46.2           +
15                  34.7           +
16                  33.3           ?
17                  26.4           ?
18                  28.8           ?
19                  28.5           ?
20                  36.8           +
21                  36.6           ?

S. marcescens       33.9
Nima
(n = 3)

E. coli             61.0
SM10 [lambda] pir

Acknowledgments

The authors wish to thank the National Science Foundation (grant # 9650766) for financial assistance with the purchase of laboratory instrumentation. Joe Wozniak provided expert technical assistance. Finally, the authors acknowledge with gratitude the gift of Escherichia coli SM10 lambda pir (pUTminiYn5Kmlux) from Dr. John Coote.

REFERENCES AND NOTES

(1.) Applegate, B. M., S. R. Kehrmeyer and G. S. Sayler. 1998. A chromosomally-based todluxCDABE whole-cell reporter for benzene, toluene, ethylbenzene and xylene (BTEX) sensing. Appl. Environ. Microbiol. 64: 2730-2735.

(2.) Bushman, F. 2002. Lateral DNA Transfer: mechanisms and consequences. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

(3.) Cornelis, G. R., A. Boland, A. P. Boyd, C. Geuijeu, M. Iriarte, C. Neyt, M. P. Sory and I. Stanier. 1998. The virulence plasmid of Yersinia, an antihost genome. Microbiol. Mol. Biol. Rev. 62: 1315-1352.

(4.) Forde, C. B., R. Parton and J. G. Coote. 1998. Bioluminescence as a reporter of intracellular survival of Bordetella bronchiseptica in murine phagocytes. Infect. Immun. 66: 3198-3207.

(5.) Gerber, N.N. 1975. Prodigiosin-like pigments. CRC Crit. Rev. Microbiol. 3: 469-485.

(6.) Haddix, P. L., E. T. Paulsen and T. E Werner. 2000. Measurement of mutation to antibiotic resistance: ampicillin resistance in Serratia marcescens. Bioscene: J. College Biol. Teaching 26: 17-21.

(7.) Haddix, P. L and T. F. Werner. 2000. Spectrophotometric assay of gene expression: Serratia marcescens pigmentation. Bioscene: J. College Biol. Teaching 26: 3-13.

(8.) Haddix, P. L. and T. F. Werner. 2002. Ampicillin sensitivity in Serratia marcescens: a model system for undergraduate genetic studies. Bioscene: J. College Biol. Teaching 28:21-32.

(9.) Meighen, E. A. 1991. Molecular biology of bacterial bioluminescence. Microbiol. Rev. 55: 123-142.

(10.) Unge, A., R. Tombolini, L. Molbak and J. K. Jansson. 1999. Simultaneous monitoring of cell number and metabolic activity of specific bacterial populations with a dual gfp-luxAB marker system. Appl. Env. Microbiol. 65: 813-821.

(11.) Photoelectric tube luminometers, including the ones used for this study, convert light to an electric current. Data are expressed as "relative" light units because cumulative luminescence is measured over a user-defined period of time. Calibration among instruments is generally inconsistent, but these can be purchased for $2,000-$3,000. Photon-counting luminometers measure photons per second, are more precise, and can be purchased for $6,000-$10,000. Photon-counting instruments may be used for this exercise, but their increased precision is not necessary.

* Corresponding author. Present address: American Water Works Service Company, Inc., 115 S. Illinois St., Belleville, IL 62220-3102. Phone: [phone]; Electronic mail: [e-mail].