guanylate cyclase (11), can be obtained from C. crescentus cell extracts. Exogenous cyclic GMP has been reported to regulate

Transcription

1 Proc. Nati. Acad. Sci. USA Vol. 73, No: 9,pp , September 1976 Microbiology Effect of 3':5'-cyclic GMP derivatives on the formation of Caulobacter surface structures (polar differentiation/synthesis of flagellin and outer membrane protein/3':5'-cyclic AMP) NURITH KURN AND LUCILLE SHAPIRO Department of Molecular Biology, Division of Biological Sciences, Albert Einstein College of Medicine, Bronx, New York Communicated by Harry Eagle, July 1, 1976 ABSTRACT Exogenous derivatives of 3':5'-cyclic GMP, 8-bromo- and N2,02-dibutyryl cyclic GMP, coordinately repress surface structure differentiation in Caulobacter crescentus. Growth in the presence of cyclic GMP derivatives resulted in the loss of flagella and pili formation and concomitant resistance to both DNA phage 4CbK and RNA phage 4CbS infection without affecting growth rate, stalk formation, and equatorial cell division. The effect of cyclic GMP derivatives was shown to be the repression of synthesis of specific structural proteins. This effect could be reversed by exogenous N6,O2-dibutyryl 3':5'-cyclic AMP, and mutants resistant to repression by cyclic GMP derivatives exhibited a pleiotropic phenotype affecting a cyclic AMP-mediated event. The sequence of microdifferentiation events that characterize the cell cycle of Caulobacter crescentus, a Gram-negative bacterium, is precisely defined both spatially and temporally and is independent of growth rate (1-4). At two distinct stages in the cell cycle, surface structures are formed at a cell pole: localized cell wall synthesis leads to the biogenesis of a polar stalk (5), and a third of the cell cycle later, a flagellum, pili, and DNA phage receptor sites are coordinately assembled at the pole opposite the existing stalk. A surface structure precursor, the flagellin subunit, was shown to be'synthesized just prior to the formation of the flagellum (6). Equatorial cell division yields a stalked cell and a motile swarmer cell (Fig. la). The structures at the pole of the incipient swarmer cell are transient and their formation appears to be coordinately controlled. (a) Nutritional deprivation after a shift from growth on glucose to growth on either lactose or galactose resulted in a cell cycle block just prior to the synthesis of flagellin and the assembly of the surface structures. Reinitiation of the cell cycle resulted in the synthesis of flagellin and the coordinate assembly of the polar structures* (b) Pleiotropic point mutants have been isolated that were unable to assemble these surface structures, although subunit synthesis was not affected (7, 8). We report here that exogenous derivatives of 3':5'-cyclic GMP, dibutyryl cyclic GMP and 8-bromo cyclic GMP, coordinately repress the formation of the polar flagellum, pili, and DNA phage 4CbK receptor sites. Cyclic GMP derivatives were shown to specifically inhibit the synthesis of flagellin and two outer membrane proteins. These results further support the contention that surface morphogenesis in Caulobacter is coordinately regulated and suggest a physiologic role for cyclic guanine nucleotide. The intracellular concentration of cyclic GMP in Caulobacter has, in fact, been shown to vary with carbon source and stage of growth of the culture.* A regulatory role for cyclic GMP was further suggested by the observation that a specific cyclic GMP receptor protein (10), as well as * N. Kurn, L. Shapiro, and N. Agabian, manuscript submitted for publication guanylate cyclase (11), can be obtained from C. crescentus cell extracts. Exogenous cyclic GMP has been reported to regulate stalk elongation in mutants of C. crescentus (12). Cyclic GMP repression of surface structure formation appears to be linked to 3':5'-cyclic AMP-mediated events, since repression was reversed by the simultaneous addition of dibutyryl cyclic AMP. Furthermore, mutants resistant to repression by derivatives of cyclic GMP were isolated both by the ability to grow on lactose at 370 in the absence of dibutyryl cyclic AMP and by selection for the maintenance of motility in the presence of cyclic GMP derivatives. MATERIALS AND METHODS "4C-Labeled reconstituted protein hydrolysate yeast profile, 1 mci/ml, was purchased from Schwarz/Mann. N6,02'-Dibutyryl 3':5'-cyclic AMP, N2,02-dibutyryl 3':5'-cyclic GMP, 8-bromo 3':5'-cyclic GMP, and N-methyl-N'-nitro-N-nitrosoguanidine were purchased from the Sigma Chemical Co. Growth Conditions. C. crescentus strain CB13 was grown at 30 in minimal medium (9) in the presence of lactose (0.5% wt/vol), mannose (0.2% wt/vol), or glucose (0.2% wt/vol). Minimal medium was prepared with deionized water produced by bacteria-free ion exchange columns. To insure low organic content, input water was first treated with iodine and bacterial filters placed before and after deionizing columns. If precaution was not taken to prevent bacterial growth on ion exchange resins, the growth rate of C. crescentus CB13 on glucose (normally 4 hr generation time) was slowed to 8 hr, cultures responded aberrantly to a carbon source shift-down, and the effect of cyclic GMP derivatives on polar differentiation was substantially decreased. Mutant Isolation. Mutant strains resistant to repression by cyclic GMP derivatives were isolated after mutagenesis with N-methyl-N'-nitro-N-nitrosoguanidine (13). Resistant mutants remained motile in the presence of 3 mm dibutyryl cyclic GMP or 0.5 mm 8-bromo cyclic GMP and were thus selected by colonial morphology on soft agar (0.1%) plates containing either cyclic GMP derivative. Motile colonies were selected on soft agar plates as described (7). Spontaneous mutants of C. crescentus CB13 able to grow on lactose as the sole carbon source at 370 appeared with a frequency of Twelve out of 12 of these mutants were shown to be simultaneously resistant to repression by derivatives of cyclic GMP. Electron Microscopy. Bacterial preparations were negatively stained with 1% phosphotungstic acid (14). Preparations were negatively stained with 2% ammonium molybdate (ph 7.0) (15) for visualization of pili and RNA phage 4Cb5 attachment. A Siemens Elmskop 1A electron microscope at a voltage of 80 kv was used.

2 3304 Microbiology: Kurn and Shapiro a b Proc. Nati. Acad. Sci. USA 73 (1976) -% _ Lr x E ^_.s_ A. SWUM -f FIG. 1. Electron micrographs of C. crescentus CB13 grown in lactose minimal medium in the absence (a) and presence (b) of 3 mm 8-bromo cyclic GMP. div, predivisional cell; st, stalked cell; sw, swarmer; f, flagellum. Bar is 1 Mm. RESULTS Effect of cyclic GMP derivatives on surface structure formation Exogenous derivatives of cyclic GMP, 8-bromo cyclic GMP and dibutyryl cyclic GMP, prevented polar structure differentiation on incipient swarmer cells. Cultures grown in the presence of lactose or mannose and cyclic GMP derivatives lost motility, as observed microscopically, 6-8 hr after cyclic nucleotide addition. Electron microscopy of the nonmotile cultures revealed the presence of all cell types, although flagella and pili were absent (Fig. lb). To improve visualization of pili, electron micrographs were taken of cultures containing an excess (multiplicity of infection = 10) of the RNA phage (Cb5. Although control cultures showed pili covered with the spherical RNA phage, greater than 90% of the nonstalked cells in the cultures grown in the presence of dibutyryl cyclic GMP lacked pili. The DNA phage #CbK has been shown to specifically attach to the C. crescentus cell pole independent of the presence of flagella or pili (6, 7, 16), but adsorption of these phage failed to occur in cultures grown in the presence of cyclic GMP derivatives. Furthermore, these cultures were resistant to both RNA phage bcb5 and DNA phage #CbK infection, as shown by streaking phage suspensions and bacterial cultures on either lactose or mannose agar plates supplemented with cyclic GMP derivatives (3 mm). Cyclic GMP derivatives had no effect on glucose-grown cultures. Although exogenous cyclic GMP derivatives repressed flagella and pili formation, as well as phage adsorption, both polar stalk formation (Fig. 1) and growth rate (Fig. 2) were unaffected. These results are consistent with previous observations that the deletion of this series of morphogenetic events by mutation does not affect other components of the Caulobacter cell cycle (7). Repression of surface structure formation appeared specific for cyclic guanine derivatives, since motility and phage sensitivity were not affected by cyclic AMP, N6,02'-dibutyryl cyclic AMP, N6-monobutyryl cyclic AMP, 8-bromo cyclic AMP, or cyclic GMP. The effective concentration of exogenous 8-bromo cyclic GMP (0.5-1 mm) was 3- to 6-fold lower than that of dibutyryl cyclic GMP. L E V. c TIME (hr) FIG. 2. The effect of exogenous cyclic nucleotides on growth of C. crescentus on lactose minimal medium. Cultures grown on glucose (0.2%) minimal medium were transferred to lactose (0.5%) minimal medium. (0) Control; (X) plus 3 mm 8-bromo cyclic GMP; (0) plus 3 mm 8-bromo cyclic GMP and 3 mm dibutyryl cyclic AMP.

3 Microbiology: Kum and Shapiro The shift of C. crescentus CB13 grown on glucose as the sole carbon source to lactose, which is utilized by inducible enzymes, resulted in a growth lag. * This lag was accompanied by an arrest of the cell cycle at the stage prior to surface structure differentiation and cell division. The cell cycle reinitiated and growth resumed concomitant with fl-galactosidase induction. The addition of dibutyryl cyclic AMP obviated the growth lag as well as the cell cycle arrest by facilitating carbon source utilization (*, 9). The 8-bromo derivative of cyclic AMP or derivatives of cyclic GMP did not mimic the stimulatory effect of dibutyryl cyclic AMP. Microscopic observation revealed, however, that, in the presence of cyclic GMP derivatives, surface structure formation remained repressed after the cell cycle and growth were reinitiated. The simultaneous addition of dibutyryl cyclic AMP and either dibutyryl or 8-bromo cyclic GMP to cultures undergoing a carbon source shift resulted in stimulation of growth (Fig. 2), and surface structure formation was not repressed. Cyclic GMP derivatives, therefore, do not interfere with the stimulatory effects of dibutyryl cyclic AMP, but the cyclic AMP derivative apparently reverses the repression of surface differentiation by cyclic GMP derivatives. Effect of exogenous cyclic GMP derivatives on protein synthesis In order to determine if repression of polar structure formation was expressed at the level of protein synthesis or coordinate assembly of precursors, cultures grown in the absence and presence of cyclic GMP derivatives were pulse-labeled with 14C-labeled amino acids. Total cell proteins were then resolved by electrophoresis on sodium dodecyl sulfate-polyacrylamide gradient slab gels.* Autoradiograms of these gels showed the profile of newly synthesized proteins (Fig. 3). Cultures grown on mannose in the presence of 8-bromo cyclic GMP lacked at least three protein bands present on autoradiograms of cultures grown on mannose alone. Dibutyryl cyclic GMP could effectively replace 8-bromo cyclic GMP in these experiments, and similar results were obtained with cultures grown on lactose in the presence and absence of cyclic GMP derivatives. Cyclic GMP derivatives apparently repressed the synthesis of the following proteins: (a) a 25,000 molecular weight protein identified as flagellin, the flagellar subunit (6), and (b) two proteins, molecular weights 73,000 and 63,000, which have been identified as outer membrane components.t The protein profile of cultures grown in the absence of cyclic nucleotide derivatives was similar to that of cultures grown in the presence of a mixture of dibutyryl cyclic AMP and 8-bromo cyclic GMP, in agreement with the microscopic observations that the repression of polar differentiation by cyclic GMP derivatives could be reversed by simultaneous addition of dibutyryl cyclic AMP. It thus appears that exogenous cyclic GMP derivatives repress the synthesis of specific structural proteins destined for polar assembly at the surface of the cell. Selection of mutants resistant to repression by cyclic GMP derivatives Mutant strains insensitive to repression by exogenous cyclic GMP derivatives were isolated by their ability to remain motile in the presence of cyclic GMP derivatives (class A), as described in Materials and Methods. Another group of cyclic GMP-resistant mutants was selected by their ability to grow on lactose at 370 (class B). Growth of wild-type cultures on lactose at 370 was possible only in the presence of 3 X 10-3 M dibutyryl cyclic t N. Agabian and B. Unger, manuscript submitted for publication. - its Proc. Nati. Acad. Sci. USA 73 (1976) 3305 A B C domumom- I I -- III FIG. 3. Autoradiograms of polyacrylamide gels of pulse-labeled C. crescentus grown in lactose minimal medium. Cultures (3 ml) were pulse-labeled for 15 min with 50 1sCi of 14C-labeled reconstituted protein hydrolysate at 300 with shaking. Resolution of labeled proteins by gel electrophoresis and autoradiography was as described [Kurn, N., Shapiro, L. & Agabian, N. (1976) J. Bacteriol, submitted for publication]. (A) Control; (B) culture grown in the presence of 3 mm 8-bromo cyclic GMP; (C) culture grown in the presence of 3 mm 8-bromo cyclic GMP and 3mM dibutyryl cyclic AMP. Marker gel (not shown) contained chymotrypsinogen, bovine serum albumin, C. crescentus RNA polymerase, and C. crescentus flagellin. I and II indicate outer membrane proteins; III indicates flagellin. AMP (Fig. 4). Mutants selected by both methods were found to have the same phenotype with respect to growth on lactose at the nonpermissive temperature and the ability to remain motile and sensitive to DNA phage OCbK infection in the presence of cyclic GMP derivatives (Table 1). Autoradiography of polyacrylamide gels of pulse-labeled cultures showed that flagellin and outer membrane proteins were synthesized in mutant cultures grown in the presence of cyclic GMP derivatives (data not shown). Electron micrographs of both mutant strains grown on lactose at 300 in the presence of cyclic GMP derivatives revealed the presence of assembled pili, and RNA phage OCb5 were observed attached to these pili. All mutant clones grown in the presence of cyclic GMP derivatives, however, were resistant to RNA phage infection. Both mutant classes retained the response to glucose-lactose shift with respect to cell cycle arrest and both could be rescued by exogenous dibutyryl cyclic AMP.

4 3306 Microbiology: Kurn and Shapiro Table 1. Proc. Natl. Acad. Sci. USA 73 (1976) Phenotype of wild-type and mutant strains of C. crescentus in the presence of exogenous 8-bromo cyclic GMP Wild-type Mutant class Al Mutant class B - 8-bromo cyclic- + 8-bromo cyclic + 8-bromo cyclic + 8-bromo cyclic Phenotype GMP GMP GMP GMP Generation time (hr)* Growth on lactose at Motilityt Flagellat Intracellular flagellin Pilit RNA phage ocb5 adsorptionl RNA phage OCb5 infection + DNA phage ocbk infection * In minimal medium with lactose as the sole carbon source. t Phase contrast microscopy. $ Electron microscopy. Autoradiograms of sodium dodecyl sulfate-polyacrylamide gels of pulse-labeled cultures. Class A mutants were selected by their ability to maintain motility in the presence of cyclic GMP derivatives. Class B mutants were selected by their ability to grow on lactose at 370 in the absence of exogenous dibutyryl cyclic AMP. DISCUSSION In this communication we report that cyclic GMP derivatives repress the formation of the polar surface structures, flagella, pili, and DNA phage infection sites of the incipient -swarmer cell prior to cell division. The subsequent formation of a stalk, as well as growth rate, were not affected by cyclic GMP derivatives. The demonstration of pleiotropic repression of polar differentiation of swarmer cells, while maintaining other cell cycle functions, supports earlier observations that the formation of these structures is coordinately controlled and not obligatory for the cell cycle (7). The target of repression by cyclic GMP derivatives was shown to be the synthesis of specific structural gene products. Cultures pulse-labeled with amino acids in the presence of cyclic GMP derivatives showed repression of flagellin synthesis, as well as the synthesis of protein components of the outer membrane. It could be argued, however,'that cyclic GMP derivatives affected the assembly of these surface structures, but that a feed-back mechanism in turn repressed the synthesis of precursor proteins. This seems unlikely, however, since synthesis of the structural proteins was not altered by a pleiotropic E CB13 +dibutyryl - camp o AE-16 CB I I I I TIME (hr) FIG. 4. C. crescentus CB13 and a mutant strain (AE16) insensitive to exogenous cyclic GMP derivatives grown in lactose (0.5%) minimal medium at 370 in the presence and absence of 3 mm dibutyryl cyclic AMP. Cultures grown on glucose (0.2%) minimal medium at 30 were transferred to lactose minimal medium. mutation that affected the coordinate assembly of these polar structures (7). Furthermore, a conditional mutant unable to assemble a flagellum at the restrictive temperature maintained production of the flagellin precursor (16). A correlation between events affected by cyclic GMP derivatives and those affected by dibutyryl cyclic AMP was demonstrated by the following observations: (a) derivatives of cyclic GMP repressed polar structure formation in cultures grown on either mannose or lactose, but not on glucose. Growth on mannose or lactose was previously shown to be temperature-sensitive. The addition of dibutyryl cyclic AMP permitted growth at the restrictive temperature and stimulated growth at the permissive temperature.* Growth on glucose, however, was unaffected by temperature or the presence of dibutyryl cyclic AMP. (b) Mutants insensitive to repression by cyclic GMP derivatives were isolated both by their ability to retain motility in the presence of the guanine cyclic nucleotides and by their ability to grow on lactose at 370 in the absence of dibutyryl cyclic AMP. (c) The repression of polar differentiation by cyclic GMP derivatives was reversed by the simultaneous addition of dibutyryl cyclic AMP. It thus appears that dibutyryl cyclic AMP interferes with repression of surface structure formation by cyclic GMP derivatives. Conversely, cyclic GMP derivatives did not interfere with dibutyryl cyclic AMP-stimulated induction of metabolic enzymes for lactose utilization. Exogenous cyclic GMP derivatives appear to repress RNA phage development in C. crescentus, independent of the repression of surface structure formation. All mutants resistant to cyclic GMP derivatives remain insensitive to RNA phage infection in the presence of these cyclic nucleotides, although they were able to assemble pili and permit RNA phage adsorption (unpublished observations). Further studies on the repression of phage development by cyclic GMP derivatives may elucidate the level at which they function. The repression of RNA phage development, together with the recent finding in E. coli of a class of relatively stable mrnas for outer membrane proteins (20,21), suggest that cyclic GMP derivatives may function in regulating the synthesis or expression of specific groups of mrna molecules. We thank Jane Fant for the electron microscopy and Ines Contreras for excellent technical assistance. This investigation was supported by grants from the National Science Foundation (GB 42545X) and the National Institutes of Health (GM 11301; GM 19100). L.S. is a Faculty

5 Microbiology: Kurn and Shapiro Research Associate of the American Cancer Society. N.K. is a recipient of a National Institutes of Health Public Health Service Fellowship (HD ). 1. Poindexter, J. S. (1964) Bacterfol. Rev. 28, Kurn, N. & Shapiro, L. (1975) in Current Topics in Cellular Regulation, eds. Horecker, B. L. & Stadtman, E. R. (Academic Press, New York), Vol. 9, pp Wood, N. B. & Shapiro, L. (1975) in Results and Problems in Cell Differentiation, eds. Reinert, J. & Holtzer, H. (Springer-Verlag, Berlin, New York), Vol. 7, pp Newton, A., Osley, M. A. & Terrana, B. (1975) in Microbiology, ed. Schlessinger, D. (American Society for Microbiology, Washington, D.C.), pp Schmidt, J. M. & Stanier, R. Y. (1966) J. Cell Biol. 28, Shapiro, L. & Maizel, J. V. (1973) J. Bacteriol. 113, Kurn, N., Ammer, A. & Shapiro, L. (1974) Proc. Natl. Acad. Sci. USA 71, Fukuda, A., Miyakawa, K. & Yoshini, 0. (1974) Proc. Jpn. Acad. 50, Proc. Nati. Acad. Scd. USA 73 (1976) Shapiro, L., Agabian-Keshishian, N., Hirsch, A. & Rosen, 0. M. (1972) Proc. Nati. Acad. Sci. USA 69, Sun, I., Shapiro, L. & Rosen, 0. M. (1975) J. Blol. Chem. 250, Sun, I., Shapiro, L. & Rosen, O. M. (1974) Blochem. Blophys. Res. Commun. 61, Schmidt, J. M. & Samuelson, G. M. (1972) J. Bacteriol. 112, Adelberg, E. A., Mandel, M. & Chan Cheng Chen, C. (1965) Biochem. Biophys. Res. Commun. 18, Shapiro, L. & Agabian-Keshishian, N. (1970) Proc. Nati. Acad. Sci. USA 67, Bendis, I. K. & Shapiro, L. (1970) J. Virol. 6, Marino, W., Ammer, S. & Shapiro, L. (1976) J. Mol. Biol., in press. 17. Sarkar, N. & Pandus, H. (1974) J. Biol. Chem. 250, Goldberg, N. D., O'Dea, R. E. & Haddox, M. K. (1973) Adv. Cyclic Nuclectide Res. 3, Bernlohr, R. W., Haddox, M. K. & Goldberg, N. D. (1974) J. Bol. Chem. 249, Lee, N. & Inouye, M. (1974) FEBS Lett. 39, Levy, S. B. (1975) Proc. Natl. Acad. Sci. USA 72,