K. Hughes and J. Roth ... chromosomal Mu d fusion as described by Maloy and Roth. (21). ..... John Roth, Ken Sanderson, Barbara Bachmann, and Don Hahn.
Vol. 169, No. 7
JOURNAL OF BACTERIOLOGY, July 1987, p. 3029-3034
0021-9193/87/073029-06$02.00/0 Copyright © 1987, American Society for Microbiology
Isolation and Characterization of Salmonella typhimurium Glyoxylate Shunt Mutantst REBECCA B. WILSON AND STANLEY R. MALOY* Department of Microbiology, University of Illinois, Urbana, Illinois 61801 Received 15 September 1986/Accepted 13 April 1987
Growth of SalmoneUa typhimurium on acetate as a sole carbon source requires expression of the glyoxylate shunt; however, the genes for the glyoxylate shunt enzymes have not been previously identified in S. typhimurium. In this study, we isolated transposon insertions in the genes for the two unique enzymes of this pathway, aceA (isocitrate lyase) and aceB (malate synthase). The aceA and aceB genes were located at 89.5 min on the S. typhimurium genetic map. Genetic linkage to nearby loci indicated that the relative gene order is purDJ metA aceB aceA. Transposon insertions in aceB were polar on aceA, suggesting that the genes form an operon transcribed from aceB to aceA. Transcriptional regulation of the aceBA operon was studied by constructing mini-Mu d(lac Kan) operon fusions. Analysis of these fusions indicated that expression of the aceBA operon is regulated at the level of transcription; the aceBA genes were induced when acetate was present and repressing carbon sources were absent. Although glucose represses expression of the aceBA operon, repression does not seem to be mediated solely by cyclic AMP-cyclic AMP receptor protein complex. Mutants with altered regulation of the aceBA operon were isolated.
the corepressor which interacts with the iclR gene product to regulate expression of the glyoxylate shunt. However, later studies showed that phosphoenolpyruvate levels do not decrease significantly during growth on acetate and hence cannot account for the induction of the glyoxylate shunt enzymes (15, 18). Thus, the actual corepressor or inducer that interacts with the iclR gene product is not yet known. In some organisms, expression of the glyoxylate shunt is related to pathogenicity (23; J. Paznokas, personal communication). Therefore, we have begun to characterize the glyoxylate shunt from the closely related bacterium, Salmonella typhimurium. Expression of the glyoxylate shunt in S. typhimurium has not been previously reported. In this paper, we describe the isolation and characterization of aceA and aceB mutations in S. typhimurium and the initial isolation of aceBA regulatory mutants.
The glyoxylate shunt is an anapleurotic pathway that allows the net accumulation of four carbon compounds during growth on two carbon substrates such as acetate or acetyl coenzyme A generated by fatty acid degradation. This accumulation is accomplished by bypassing the two C02evolving steps of the tricarboxylic acid cycle (5, 13). In Escherichia coli, the two unique enzymes of the glyoxylate shunt, isocitrate lyase (aceA) and malate synthase (aceB), are induced during growth on acetate or fatty acids. In E. coli, the aceA and aceB genes are in an operon that is transcriptionally regulated by two repressors encoded by the iclR and fadR genes (20). In addition to regulating the aceBA operon, the fadR repressor also controls expression of the genes for fatty acid degradation; expression of the fad genes is induced by long-chain fatty acids (>C12) (reviewed in reference 24). Although the glyoxylate shunt in E. coli has been studied for many years, the mechanism of genetic regulation remains unclear. Genetic studies suggest that acetate or acetyl coenzyme A is not the inducer of the glyoxylate shunt (14). Other potential regulatory mechanisms have been proposed based on the physiology of growth on acetate. When cells grown on another carbon source are transferred to media with acetate as the sole carbon source, dicarboxylic acids from the tricarboxylic acid cycle are depleted for biosynthetic uses. This depletion has at least two consequences: there is inadequate oxaloacetate to combine with acetyl coenzyme A to run the tricarboxylic acid cycle, and there is a consequential drop in the intracellular concentration of other intermediates which must be derived from the tricarboxylic acid cycle. Decreasing the concentration of an intermediate which also functions as a corepressor of the glyoxylate shunt could allow induction of the glyoxylate shunt enzymes. Kornberg (14) suggested that phosphoenolpyruvate may be
MATERIALS AND METHODS Bacterial strains. The genotypes of the strains used in this study are shown in Table 1. All S. typhimurium strains were derived from LT2. Media and growth conditions. Nutrient broth (0.8%; Difco Laboratories, Detroit, Mich.) with 0.5% NaCl added was used for rich medium. The minimal medium used was NCE (2). Carbon sources were added to minimal medium at the following final concentrations: 0.4% potassium acetate, 0.6% sodium succinate, and 5 mM potassium oleate. When oleate was added to media, the detergent Brij 58 was also added at 0.5%. Amino acid supplements were added at the concentration listed in Davis et al. (8). Antibiotics were added at the following final concentrations: kanamycin sulfate, 125 ,ug/ml in minimal medium or 50 ,ug/ml in rich medium tetracycline hydrochloride, 10 ,ug/ml in minimal medium or 20 pug/ml in rich medium; sodium ampicillin, 15 ,ug/ml in minimal medium or 30 ,ug/ml in rich medium; and streptomycin sulfate, 2 mg/ml in minimal medium with 1% nutrient broth added. 5-Bromo-4-chloro-3-indolyl-p-D-galactoside (X-gal) was dissolved in N,N-dimethylformamide at a concentration of 20
* Corresponding author. t This paper is dedicated to the memory of our late friend and colleague William Nunn. 3029
3030
WILSON AND MALOY
J. BACTERIOL.
TABLE 1. Bacterial strains used Organism and strain
S. typhimurium LT2 TR5670 TR5878
TT627 TT10289 PP1002 PP1037 SA2601 MS226 MS229 MS505 MS510 MS1309 MS1311 MS1360 MS1361 MS1380 MS1381 MS1393 MS1398 MS1517 MS1519 E. coli K-12
JC1553/KFL10
Genotypea
Source
Prototrophic metA53 rpsLl hsdLT6 hsdS29 (rLT- mLT+ r,- m,+) ilv452 metA22 trp,2 metE551 xyl404 fla-66 rpsLI20 HI-b nml H2-e,n,X (Fels2-) pyrC7 rpsLlIF'tsll4 lac+ zzf-20::Tn1O hisD9953::Mu dJ his-9949::Mu d-1 cya: :TnJO trpB223 crp-773::Tn1O trpB223 purD130
aceA101::TnJO aceB102: :TnlO metA53 aceA102::TnlO rpsLI purHJ391 aceA102::TnJO aceA112::Mu dJ aceB113::Mu dJ aceA112::Mu dJ/F'110 ace' zzf-20::TnlO aceBJ13::Mu dJ/F'110 ace' zzf-20::TnlO aceA12::Mu dJ cya::TnlO aceBJ13::Mu dJ crp-773::TnlO aceAl12::Mu dJ aceB102::TnJO aceB113::Mu dJ aceAJOl::TnlO putA1019::Mu dA(bla;:Tn5) cya::TnlO putA1019: :Mu dA(bla: :TnS) crp-773: :TnJO
Prototrophic leuB6 hisGl argG6 metBlz lacYl gal-6 malAl xyl-7 mtl-2 rpsL104 tonA2 tsx-l supE44 recAl lambda-/F' 110
J. Roth Davis et al. (8) J. Roth
Davis et al. (O K. Hughes and J. Roth P. Postma P. Postma M. Demerec via SGSCb This study This study This study This study This study This study This study This study This study This study This study This study D. Hahn D. Hahn CGSCC K. Low via CGSCC
Genetic nomenclature is as described by Sanderson and Roth (25). Nomenclature for Mu d insertions is described in Materials and Methods. b Obtained from K. Sanderson at the Salmonella Genetic Stock Center (SGSC), Department of Biology, University of Calgary, Calgary, Alberta, Canada. c Obtained from B. Bachmann at the E. coli Genetic Stock Center (CGSC), Yale University, New Haven, Conn.
a
mg/ml and then added to media to yield a final concentration of 20 ,g/ml. 3', 5'-Cyclic AMP (cAMP) was added to media at a final concentration of 5 mM. The carbon sources, antibiotics, and supplements were all obtained from Sigma Chemical Co., St. Louis, Mo. Solid medium contained 1.5% Bacto-Agar (Difco). Genetic techniques. The high-frequency generalized transducing bacteriophage P22 HT105/1 int-201 was used for all transductions (26). Phage lysates were prepared as described by Davis et al. (8). Transductions were performed by infecting cells with phage at a multiplicity of about 1 PFU per cell. Usually phage and bacteria were mixed directly on selective medium. When Kanr and Strr colonies were selected, phage and bacteria were spread on nonselective medium, incubated for 4 to 5 h at 37°C, and then replica plated onto a medium containing the antibiotic. Transductions were purified, and phage-free clones were isolated by streaking them nonselectively on green indicator plates (6). The phage-free colonies were then checked for P22 sensitivity by crossstreaking them against P22 H5 (a clear-plaque mutant) (8). Conjugational matings were performed directly on selective plates as described previously (22). Donor strains containing temperature-sensitive F-prime plasmids were grown at 300C. Hfr strains were formed by recombination between the lac homology on F'ts114 lac+ and the lac region on the chromosomal Mu d fusion as described by Maloy and Roth (21). Operon fusions. A simplified nomenclature suggested by K. Hughes and J. Roth (in press) is used to describe the Mu d operon fusion vectors used in this study. Mu dl(Amp lac) is the original operon fusion vector constructed by Casadaban and Cohen (4). It carries all essential Mu transposition
functions, and expression of these functions is controlled by a temperature-sensitive repressor (cts). This vector is simply designated Mu dl. Mu d11734 is a derivative of Mu dl which has been deleted for transposition functions and carries Kanr instead of Ampr (5). Mu d11734 is designated Mu dJ. Mu dl-8 is a derivative of Mu dl that has amber mutations in the Mu transposition functions (12). Mu dl-8 is designated Mu dA. Isolation of aceBA mutants. Stable ace::Mu dJ fusions were isolated in two ways. Random Mu dJ insertions were obtained by transitory trans-acting transposition (Hughes and Roth, in press). LT2 was transduced to Kanr with phage grown on strain TT10289. This strain carries two Mu d insertions in the his operon, Mu dJ (Kan') and Mu dl (Amp'). Because of size limitations, P22 cannot package both Mu d phage into a single transducing particle. Mu dJ does not carry the Mu transposition functions, but Mu dl does. When a transducing fragment carries the entire Mu dJ genome and the portion of the Mu dl genome that encodes transposase, the transposase gene can be transiently expressed to provide transposase which can act in trans on the Mu dJ genome, allowing Mu dJ to transpose to new sites in the chromosome. The linear transducing fragment is rapidly degraded by cellular nucleases, eliminating transposase expression. Thus, once the Mu dJ is integrated into the chromosome, it is stable. Homologous recombination of the Mu dJ transducing fragment into the chromosome yields His- Kanr colonies, while transposition usually yields His' Kanr progeny. To eliminate inheritance of the Mu dJ by homologous recombination, transductants were selected on minimal medium containing kanamycin sulfate. The fHis+ Kanr transductants were replicated onto X-gal, acetate, oleate, succinate, and sodium ampicillin plates. X-gal+ Ace- Ole-
VOL. 169, 1987
Suc+ Amps Kanr colonies were picked as potential aceBA::Mu dJ fusions. The fusions were then mapped with respect to metA, and enzymatic assays were performed to confirm the genotypes of the mutants. We also isolated aceBA::Mu dJ insertions by localized mutagenesis (8, 11). To accomplish this, strain TR5670 (metA) was transduced with a P22 phage stock grown on a pool of colonies with random Mu dJ insertions, and Met+ His+ Kanr transductants were selected on minimal medium containing kanamycin sulfate. About 13% of these insertions had the phenotype expected for aceBA mutants (Ace- OleSuc+ Amps). TnJQ insertions in and near the aceBA genes were also isolated by localized mutagenesis. Strain TR5670 (metA) was transduced to Met+ Tetr with phage grown on a pool of random TnlO insertions in LT2. The transductants were screened and characterized as described above. Mutant derivatives of the ace::Mu dJ fusions that were defective for cya or crp were constructed by transduction of the fusions with phage grown on PP1002 (cya::TnlO) or PP1037 (crp::TnlO). Transductants were selected on MacConkey ribose glycerol plates (1) that contained tetracycline hydrochloride. White Tetr colonies were chosen as cya or crp mutapts. Although the aceBA::Mu dJ fusions were X-gal+, the level of expression of the lac genes was insufficient to allow growth on lactose as a sole carbon source (Lac-). Regulatory mutants were isolated by selecting for increased expressions of the lac operon from aceBA::Mu dJ operon fusions. Samples (0.1 ml) of Qvermnght cultures of ace::Mu dJ fusion strains were spread on minimal lactose plates, and a few crystals of nitrosoguanidine were placed in the center of the plate. Lac+ colonies that arose were restreaked on minimal lactose plates and tested for constitutive 3-galactosidase expression.
Complementation with F'ace+. F'110 carries the region of the E. coli K-12 chromosome between polA and malB (17), including the metA and aceBA+ genes (20). To transfer F'110 to S. typhimurium, the E. coli donor JC15533/KFL10 (F'110) was mated with the S. typhimurium recipient TR5878 (r- m+ metA metE), with selection for MetA+ and counterselection against the auxotrophic mutations in the donor. Vitamin B12 was included in the medium to suppress the metE mutation in TR5878 (13). The resulting exconjugants (TR5877/F'110) were transduced to Tetr with phage P22 grown on TR627, yielding TR5877/F'110 zzf::TnJO. F'110 zzf:-TnJO was then mated into other S. typhimurium strains by selection for Tetr and counterselection against the auxotrophic mutations in the donor. Enzyme assays. Cells were grown to mid-log phase (100 to 120 Klett units) in 100 ml of minimal medium, and crude extracts were prepared in a French press (19). Specific activities of isocitrate lyase and malate synthase were determined by using spectrophotometric assays previously described (19). Isocitrate lyase activity was measured by determining the rate of glyoxylate formation, and malate synthase activity was measured by determining the rate of hydrolysis of acetyl coenzyme A. Protein concentrations of the crude extracts were determined by a modified Bradford assay (Bio-Rad Laboratories, Richmond, Calif.) with bovine serum albumin as a standard. ,-Galactosidase activity was measured as described by Miller (22) by using the chloroform-sodium dodecyl sulfate permeabilization procedure. P-galactosidase activity was expressed as nanomoles per minute per optical density unit at 650 nm.
S.
TYPHIMURIUM GLYOXYLATE SHUNT MUTANTS
*
