Proc. Natl. Acad. Sci. USA Vol. 77, No. 10, pp. 5799-5801, October 1980 Biochemistry

Catabolite repression in mutants lacking cyclic AMP receptor protein ( effect/ termination/) CHANTAL GUIDI-RONTANI*, ANTOINE DANCHINt, AND AGNES ULLMANN* *Dpartement de Biochimie et Genetique Moleculaire, Institut Pasteur, 28, rue du Docteur Roux, 75015 Paris, France; and tInstitut de Biologie Physico-Chimique, 13, rue Pierre et Marie Curie, 75005 Paris, France Communicated by Roger Y. Stanier, July 10, 1980

ABSTRACT Pleiotropic carbohydrate-positive pseudo- repression. This result indicates strongly that catabolite re- revertants have been isolated from a specific class of rho-crp pression can take place in strains lacking functional CAP pro- double mutants of Escherichia coli carrying both defective vided that they are also deficient in rho transcription termination protein, rho, and cyclic AMP receptor protein. protein. The modulation of catabolite repression of ft-galac- tosidase, amylomaltase, and tryptophanase has been studied MATERIALS AND METHODS in the pseudorevertants. It has been found that these mutants Strains and Growth Media. The exhibit catabolite repression. Because catabolite-sensitive op- following E. coli K-12 erons can be expressed in the absence of functional cyclic AMP strains were used: PP7811 (F-, argH, his); PP7812, isogenic receptor protein, this would suggest on the one hand that the with PP7811 except for the rho tsl5 allele (8); PP7810 (F-, ilvA, cyclic AMP-receptor protein complex is not the exclusive me- argH, his); and G61 (Hfr G6, aroB, his). All mutants derived diator of catabolite repression anKon the other hand that rho from these strains are described in the text. Strains were grown might be involved in the regulation of catabolite-sensitive op- in LB or 63 minimal medium (9) supplemented with the re- erons. quired amino acids, thiamine, and glucose or succinate as car- bon sources. It is generally believed that, in Escherichia coil, adenosine 3',- 5'-cyclic monophosphate (cAMP) and its receptor protein (CAP) Enzymatic Assays. f-Galactosidase (EC 3.2.1.23) was as- are specifically required for sayed according to Pardee et al. (10), amylomaltase (EC transcription initiation of cata- 2.4.1.25) according to Schwartz (11), and tryptophanase (EC bolite-sensitive . Mutants deficient for adenylate cyclase to Suelter et al. in (cya) and for CAP (crp) are unable to metabolize a great 4.1.99.1) according (12) toluenized bacterial number of carbohydrates, and the expression of catabolite- suspensions. Galactokinase (EC 2.7.1.6) was assayed as described sensitive operons is reduced. by Wetekam et al. (13) in sonicated bacterial extracts. One unit severely Several lines of evidence is the amount of that converts 1 nmol of substrate per suggested, however, that cAMP is not the unique regulator of min at 280C. catabolite repression (1, 2). Furthermore, expression of the lactose and cAMP Assay. Stationary-phase cultures were diluted to ob- arabinose operons independent of cAMP and CAP tain 300,gg (dry weight) of bacteria per ml. The bacterial sus- has been described (3, 4) and shown to be due to specific pensions were heated for 5 min at 100'C, and the total amount mutations in the regulatory regions of the operons. Another of cAMP was determined by a standard radioimmunoassay. category of mutants (alt mutants affecting the a subunit of Reagents. They were obtained from the following com- RNA polymerase) (5, 6) has been obtained as Ara+ revertants panies: ['4C] from Amersham, cAMP reagent kit from from a cya strain synthesizing reduced levels of cAMP. These Institut Pasteur Production, isopropyl-f3-D-thiogalactoside, mutants partially overcome cAMP deficiency for ,B-galacto- D-fucose, and cAMP from Sigma, and all other chemicals from sidase synthesis at low temperature. However, pleiotropic de- Merck. repression of catabolite-sensitive operons in the absence of functional CAP has never been described. This kind of study would require, in the first place, obtaining pleiotropic carbo- RESULTS hydrate-positive revertants (Lac+, Mal+ Ara+, Tna+) from crp Isolation of rho-crp Mutants. Our former studies have mutants. Such mutants have not, as yet, been found. The ra- shown that rho-mediated natural polarity can be overcome by tionale to search for these kinds of mutants was provided by our the cAMP-CAP complex (7). The growth thermosensitivity of recent results (7) that show that the cAMP-CAP complex is strains carrying the rho tsl5 allele is usually ascribed to the involved in transcription termination, suggesting a functional absence of transcription termination (8). We reasoned that, if relationship between this complex and the transcription ter- cAMP-CAP is relieving spontaneous termination, mutants mination protein rho. Therefore, we argued that if we could defective in the production of this complex might be protected obtain appropriate mutants deficient in both rho and CAP against the deleterious effects of the rho ts15 mutation. We activity, these might be good candidates for selecting pseudo- therefore selected growing colonies on McConkey maltose revertants. exhibiting catabolite repression in the absence of plates at nonpermissive temperature (44°C) from strain functional cAMP-CAP complex. In the present paper we show PP7812. When 108 colonies were incubated on such a plate, that a specific class of double mutants, rho-crp (displaying about 50 growing clones were selected. Among these, 10 were carbohydrate-negative character), can give rise to pleiotropic phenotypically Mal-. We analyzed in more detail 192 such carbohydrate-positive pseudorevertants exhibiting catabolite Mal- mutants. All were pleiotropic carbohydrate negative and five among them were characterized as cya mutants. All other The publication costs of this article were defrayed in part by page clones behaved like CAP- mutants. Several were mapped and charge payment. This article must therefore be hereby marked "ad- vertisement" in accordance with 18 U. S. C. §1734 solely to indicate Abbreviations: cAMP, adenosine 3',5'-cyclic monophosphate; CAP, this fact. cAMP receptor protein. 5799 5800 Biochemistry: Guidi-Rontani et al. Proc. Natl. Acad. Sci. USA 77 (1980) were found to be mutated at or near the crp locus. One such in most of the strains. The main feature of these experiments mutant, RCC2, was kept for further study and was used as the appeared when glucose was used as a carbon source: three parent strain in the present work. strains were sensitive to catabolite repression exerted by glu- Isolation of Pleiotropic Carbohydrate-Positive Pseudo- cose-i.e., the differential rates of enzyme synthesis were re- revertants from rho-crp Mutants. An overnight culture of duced (to different extents depending upon the strains) com- strain RCC2 grown in LB medium at 41'C was centrifuged, pared to those obtained in succinate. In one strain, CGR16, resuspended in 63 minimal medium, and plated on maltose glucose did not seem to exert a catabolite repression effect be- minimal agar plates supplemented with eosin/methylene blue cause enzyme levels were not significantly different in glucose (109 bacteria per plate). After 48 hr of incubation at 41'C, compared to succinate. growing clones appeared at a frequency of about 10-7. Ten In strains carrying wild-type CAP, cAMP antagonized to independent clones were tested for growth on lactose or arab- some extent catabolite repression exerted by glucose. As can be inose as carbon source. All exhibited Lac+ Ara+ Mal+ pheno- seen in Table 1, cAMP stimulated ,B-galactosidase synthesis in type. Four of such pleiotropic carbohydrate-positive pseudo- the two parental strains (PP7811 and PP7812), whereas in strain revertants will be described herein. RCC2 and its derivatives it had no effect on enzyme synthesis, Functional Analysis of Mutants. As will be shown below, as might be expected for CAP-defective mutants. the pseudorevertants retained their crp mutation. According Genetic Analysis of Mutants. Our conclusions are strongly to the usual model, it therefore becomes difficult to visualize dependent on the genetic characterization of the mutants- how they can express catabolite-sensitive operons in the absence namely, their crp and rho ts15 characters. They were controlled of functional CAP. We therefore measured the level of ex- as follows. pression of (3-galactosidase, amylomaltase, and tryptophanase, (i) crp mutation: Absence of cAMP stimulation of f3-galac- known to be sensitive to catabolite repression, under two ex- tosidase synthesis is circumstantial evidence suggesting that treme conditions: growth in the presence of glucose (condition CAP is lacking in the mutants. We have also measured cAMP of severe catabolite repression) and in the presence of succinate levels in the mutants. Indeed, it has been noted (14) that crp (catabolite-sensitive operons are fully derepressed). mutants overproduce cAMP by a factor of 3-10; this was con- Table 1 compares the original strains, PP7811 (rho+), PP7812 firmed in all our mutants (Table 2). (rho tsl5), and RCC2 (rho tsl5 crp), with the spontaneous P1 transduction into an aroB strain for AroB+ character re- carbohydrate-positive derivatives of RCC2. As expected, none sulted in 24-40% pleiotropic carbohydrate-negative bacteria. of the three catabolite-sensitive was expressed in strain This strongly supports our assumption that the crp character RCC2. In the carbohydrate-positive derivatives, these enzymes was still present in the mutants. Finally, with a sensitive and were expressed at various levels in succinate medium, attaining specific radioimmunoassay (15), no CAP could be detected in wild-type levels for tryptophanase expression in all strains and the extracts of our mutants. levels for and amylomaltase (ii) rho tslS mutation: We have checked the presence of the remarkably high f3-galactosidase rho ts15 allele by two independent criteria: measurement of Table 1. Expression of fl-galactosidase, amylomaltase, and relief of polarity in the gal and mapping of the mutation tryptophanase in pleiotropic carbohydrate-positive pseudo- with respect to the ilv locus. Rho factor is thought to account revertants of crp-rho mutants and in the parental strains for the major polar effects in polycistronic operons. We have previously shown that in a rho-defective mutant the rate of Enzyme synthesis, units/mg synthesis of the distal of the gal operon, galactokinase, is Carbon ,B-Galacto- Amylo- Trypto- strongly elevated (7). Table 3 shows that galactokinase levels Strain source cAMP sidase maltase phanase were very high in the mutants compared to the parent PP7811 PP7811 Succinate - 11,360 153 520 and similar to the level obtained in strain PP7812 harboring the Glucose - 2,860 <0.5 <4 rho ts15 allele. Glucose + 6,070 36 480 We have also transduced an Ilv- strain for Ilv+ character by PP7812 Succinate - 10,150 136 360 using P1 grown on our mutants. In all cases, we have obtained Glucose - 2,440 <0.5 <4 thermosensitive transductants, albeit at rather low frequencies Glucose + 6,880 15 30 probably because of the deleterious effect of the presence of RCC2 Succinate - 90 <0.5 <4 the rho ts allele. These results strongly suggest that our mutants Glucose - 180 <0.5 <4 have kept their rho tsl5 mutations. Glucose + 180 ND ND CGR11 Succinate - 5,720 121 280 Glucose - 2,300 <0.5 20 Table 2. cAMP levels in parental and mutant strains: Glucose + 2,400 ND ND Transduction mapping of crp CGR12 Succinate - 7,417 70 230 Total cAMP Glucose - 1,723 20 14 concentration, Glucose + 2,000 ND ND nmol/mg dry Frequency, CGR15 Succinate - 14,330 47 320 Strain weight bacteria % Glucose - 3,990 25 47 Glucose + 4,190 ND ND PP7811 8 CGR16 Succinate - 2,890 83 280 PP7812 8 Glucose - 2,850 80 124 RCC2 60 24 Glucose + 2,890 ND ND CGR11 53 36 CGR12 63 40 Cultures of strains were separately induced for f3-galactosidase (1 mM isopropylthiogalactoside), amylomaltase (10 mM maltose), and CGR15 22 36 tryptophanase (10 mM tryptophan). cAMP (5 mM) was added con- CGR16 20 26 comitantly with the . After about six generations of growth For measurements of cAMP concentration, strains were grown in at 410C (strain PP7812 was grown at 37°C), the enzyme activities were glucose/minimal medium at 410C (strain PP7812 was.grown at 37°C). determined. Results are expressed in units of enzyme per mg (dry For mapping of crp, AroB+ transductants were selected at 370C and weight) of bacteria. ND, not determined. tested for Maltose- and Lactose- characters. Biochemistry: Guidi-Rontani et al. Proc. Natl. Acad. Sci. USA 77 (1980) 5801

Table 3. Expression of galactokinase in parental and mutant one could hope to find pleiotropic carbohydrate-positive strains: Transduction mapping of rho pseudorevertants from our rho ts-crp background. We have Galactokinase, Frequency, shown this to be the case. The most striking feature of our results Strain units/mg % is the finding that regulation of catabolite repression can take place in a crp background (i.e., in the absence of a functional PP7811 6 cAMP-CAP complex). PP7812 42 The simplest model that might account for our results is to RCC2 38 24 suppose that cAMP-activated CAP interacts (directly or indi- CGR11 40 20 rectly) with RNA polymerase at specific sites located near or CGR12 30 16 at the . This functional interrelationship persists during CGR15 44 44 the CGR16 progression of RNA polymerase. The presence of CAP 29 72 would therefore prevent a premature action of rho. If in cata- Strains growing in glucose/minimal medium at 370C were induced bolite-sensitive operons transcription termination can occur at by 2 mM fucose for about seven generations. The results are expressed an early stage in the absence of CAP, the operon will not be in units ofgalactokinase per mg (dry weight) ofbacteria. For mapping transcribed unless rho ofrho, Ilvt transductants were selected at 30WC and tested for ther- protein is not functional. It seems, mosensitivity by plating at 300C and 410C. however, that a double mutation crp-rho is not sufficient to allow efficient transcription of catabolite-sensitive operons. Our These results leave little doubt, if any, that the mutants ex- finding that all pleiotropic carbohydrate-positive revertants hibiting pleiotropic carbohydrate-positive phenotype are de- acquired resistance to rifampicin suggests that an alteration in ficient in functional CAP and rho proteins. The question that RNA polymerase or in an additional transcription factor can therefore arises is what additional mutation(s) occurred that generate the formation of an "open" complex needed for the enables the expression of catabolite-sensitive operons. The first initiation of the transcript. approach we used to detect specific alterations was to search Our results, showing a regulation of catabolite repression in for resistance to drugs that affect the transcription machinery. strains lacking CAP, strengthen our former conclusion that the We indeed found that all carbohydrate-positive revertants cAMP-CAP complex contributes only partially (and perhaps became resistant to rifampicin (20-504,g/ml depending upon indirectly) to catabolite repression and that this complex might the mutants). Preliminary mapping data performed with have a much more general function than the one that is gen- mutants CGR11 and CGR16 were compatible with a mutation erally assumed. in the rpoB gene; they yielded 85% cotransduction frequency with the proximal purH marker. This result strongly suggests We thank Nicole Guiso for performing the radioimmunoassays of that a modification of the structure of RNA polymerase might CAP. This work was supported by grants from the Centre National de have occurred. la Recherche Scientifique (Laboratoire associe no. 270 and Groupe de Recherche no. 18) and the Fondation pour la Recherche M6dicale. DISCUSSION 1. Dessein, A., Schwartz, M. & Ullmann, A. (1978) Mol. Gen. Genet. Glucose or other rapidly metabolizable substrates in growth 162, 83-87. media elicit a severe inhibition of catabolic enzymes, a phe- 2. Wanner, B. L., Kodaira, R. & Neidhart, F. C. (1978) J. Bacteriol. nomenon referred to as catabolite repression (16). According 136,947-954. to the generally accepted model, the CAP-CAMP complex 3. Arditti, R., Grodzicker, T. & Beckwith, J. (1973) J. Bacteriol. 114, exerts a positive control at the level of transcription initiation 652-655. of 4. Colome, J., Wilcox, G. & Englesberg, E. (1977) J. Bacteriol. 129, catabolite-sensitive operons. Several lines of evidence already 948-958. suggested that cAMP might not be the unique mediator of ca- 5. Silverstone, A. E., Goman, M. & Scaife, G. (1972) Mol. Gen. tabolite repression (1, 2). Moreover, our results on the in- Genet. 118,223-234. volvement of the CAP-cAMP complex in transcription ter- 6. Travers, A. A., Buckland, R., Goman, M., Le Grice, S. S. G. & mination have cast some doubt on the exclusive role that is Scaife, J. G. (1978) Nature (London) 273,354-358. ascribed to this complex at transcription initiation. In order to 7. Ullmann, A., Joseph, E. & Danchin, A. (1979) Proc. Natl. Acad. establish clearly that regulation of catabolite repression can take Sci. USA 76,3194-3197. place in the absence of CAP-cAMP, it was important to show 8. Das, A., Court, D. & Adhya, S. (1976) Proc. Natl. Acad. Sci. USA that strains lacking CAP would exhibit this regulation. So far, 73, 1959-1963. no mutants have been found that express catabolite-sensitive 9. Miller, J. H. (1974) Experiments in (Cold operons in a crp background. In fact, among more than Spring Harbor Laboratory, Cold Spring Harbor, NY). 1000 10. Pardee, A. B., Jacob, F. & Monod, J. (1959) J. Mol. Biol. 1 Mal+ pseudorevertants obtained at a frequency of 10-8 from 165-178. various crp mutants, none was found to be Lac+ (data not 11. Schwartz, M. (1967) Ann. Inst. Pasteur 112,673-700. shown). Our previous finding that the CAP-cAMP complex acts 12. Suelter, C. M., Wang, J. & Snell, E. E. (1976) FEBS Lett. 66, like an antipolar device probably by interfering with premature 230-232. transcription termination uncovered a functional relationship 13. Wetekam, W., Staack, K. & Ehring, R. (1972) Mol. Gen. Genet. between this complex and the transcription termination protein 116,258-276. rho. This led us to conjecture that one might find crp mutants 14. Fraser, A. D. & Yamazaki, H. (1978) Can. J. Biochem. 56, among thermoresistant clones isolated from a rho ts strain. We 849-852. found that this was the case, provided that no other mutation 15. Guiso, N. & Blazy, B. (1980) Biochem. Biophys. Res. Commun. such as rpsL interfered with polarity (ref. 94,278-283. 17 and unpublished 16. Magasanik, B. (1961) Cold Spring Harbor Symp. Quant. Biol. data). This result led us to infer that cAMP-CAP action at the 26,249-256. promoter of catabolite-sensitive operons might reflect an in- 17. Petersen,.H. U., Joseph, E., Ullmann, A. & Danchin, A. (1978) terference with a premature action of rho. If this were the case, J. Bacteriol. 135,453-459.