Mol Gen Genet (1984) 193:210-213 © Springer-Verlag 1984

Autoregulation of the rho Gene of Escherichia coil K-12 Hsiang-fu Kung 1, Eva Bekesi 1, Sonia K. Guterman 2* John E. Gray 3**, Lisa Traub 3, and David H. Calhoun 3 1 Department of Molecular Genetics, Hoffman-La Roche Inc., Nutley, New Jersey 07110, USA 2 Biological Sciences Center, Boston University, 2 Cummington Street, Boston, Massachusetts 02215, USA 3 Department of Microbiology, Mount Sinai School of Medicine, Fifth Avenue and 10ffh Street, New York, New York 10029, USA

Summary. It has previously been proposed, based on indirect evidence, that the Rho protein may control the expression of the rho gene. Using an in vitro system for the transcription and translation of the rho gene cloned into plasmid pBR322, we tested this hypothesis directly by monitoring the effect in vitro of excess or limiting Rho protein. The addition of purified Rho protein suppresses Rho synthesis in vitro. The addition of antibody to Rho specifically stimulates Rho synthesis in vitro. The stimulation of Rho factor synthesis by antibody to Rho is reversed by Rho protein. Rho factor purified from a strain with a mutationally altered rho gene (rho-ll5) does not suppress Rho synthesis in vitro. These results provide convincing evidence that the rho gene is subject to autoregulation.

Introduction

Transcription termination factor Rho, the product of the rho gene, was originally identified (Roberts 1969) as a protein required for specific termination of transcription and release of RNA during synthesis in vitro. It is now recognized that Rho factor participates in various termination events, including some or all attenuation (Yanofsky 1981), termination of transcription at the end of a gene or operon (Platt 1981), polarity due to nonsense or insertion mutations (Richardson et al. 1975; Reyes et al. 1976), antiterruination mediated by the phage lambda gene N product (Adhya et al. 1974), and/or L factor, the nusA gene product (Kung et al. 1975; Greenblatt et al. 1981). Mutations in the rho gene have been described in E. coli (Ratner 1976; Imai and Shigesada 1978) and Salmonella typhimurium (Housley et al. 1981) that result in the production of elevated levels of a mutationally altered Rho protein. S. typhimurium strains that are diploid for the wild type gene produce levels of Rho protein, as judged by poly(C)dependent ATPase activity, equal to that of a haploid strain (Housley et al. 1981). These findings are consistent with and have led to the suggestion that the rho gene is subject to autoregulation (Ratner 1976; Imai and Shigesada 1978; * Present Address." Biotechnical International Inc., 85 Bolton

Street, Cambridge, Massachusetts 02140, USA ** Present Address: E.I. dn Pont de Nemours & Company, Experi-

mental Station, E328/264A, Wilmington, Delaware 19898, USA Offprint requests to: D.H. Calhoun

Housley et al. 1981). We previously cloned an 8 kilobase (kb) HindIII segment containing the rho gene from 2dilv-rho phage to plasmid pBR322 (Gray et al. 1981), and further localized the gene to a region of about 2 kb defined by two KpnI sites (Calhoun et al., accompanying article). In the studies reported here we used a DNA dependent in vitro coupled transcription and translation system to test for autoregulation of the rho gene. Materials and Methods Bacteria and Plasmids. The strain used to prepare the fractionated extracts for protein synthesis was E. coli K-12 Z19i q (Kung et al. 1979). The plasmid with the rho gene

was pJG32, a pBR322 derivative (Gray et al. 1981). This plasmid contains an 8 kb insert that includes the rho structural gene and continguous DNA sequences required for expression (Calhoun et al., accompanying article). Plasmid Preparation and in vitro Synthesis. The plasmid

DNA was prepared from cells following chloramphenicol amplification and CsC1 centrifugation as previously described (Gray et al. 1981). The preparation of the extracts and conditions for in vitro protein synthesis have been described (Kung et al. 1975, 1979). The system (35 ~tl) contained 15 mM Tris-acetate, pH 8.2, 11 mM sodium dimethylglutarate, 35 mM ammonium acetate, 65 mM potassium acetate, 10 mM magnesium acetate, 0.8 mM spermidine hydrochloride, 2.4 mM dithiothreitol, 0.93 mM each of UTP, CTP, and GTP, 3.0 mM ATP, 24 mM phosphoenolypyruvate, 0.2 l~g of pyruvate kinase, 0.05 mM ppGpp, 0.7 mM 3'5'-cAMP, 15 p.M calcium leucovorin, 12.5 I~g of E. coli tRNA, 0.63 mg of poly(ethylene glycol)6000, approximately 6 gg of DNA template, 0.4 mM isopropyl-thio-fl-D-galactopyranoside, 0.112 mM each of every amino acid, 65 Ci [35S]-methionine, 1.2 A260 units of ammonium choride washed ribosomes, ribosomal wash (50 ~tg of protein), a 0.25 M DEAE salt eluate (110 gg of protein), a 1 M DEAE salt eluate (7.5 gg of protein), EF-Tu (3 gg) and Ehrlich's ascites S-100 (26 gg). Where indicated, the antibody (1 ~tl) was added to the incubation mixture before the addition of DNA template. The reaction was carried out at 37°C for 90 min. The in vitro synthesis products were analyzed on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (Laemmli 1970) with 12.5% uniform polyacrylamide gels. The stacking gel was 5% acrylamide. Samples (3 gl) were dissolved in 10 gl of sample buffer and boiled as indi-

211 cated. Electrophoresis was carried out either overnight at a constant voltage of 30 V or for 3 4 h at 100 V. Gels were stained with Coomassie brilliant blue, destained, treated with Enhance (NEN) dried and exposed for autoradiography on X-Omat R film (Kodak). Molecular weight standards used were the following: phosphorylase B, 94 K D (kilodaltons); bovine serum albumin, 68 K D ; ovalbumin, 45 KD; carbonic anhydrase, 30 KD; soybean trypsin inhibitor, 21 K D ; and lysozyme, 14 KD.

B-GAL

RHO

Preparation o f Rho and Antibody to Rho. The Rho factor

was purified as previously described (Kent and Guterman, 1981). Rabbit antibody to the purified protein was purified by ammonium sulfate fractionation and DEAE-Sephadex column chromatography. Serum from immunized rabbits (approximately 41 ml/rabbit) was brought to 50% saturation with ammonium sulfate, the precipitate was rinsed in 40% saturated ammonium sulfate and centrifuged. The pellet was dissolved in 0.023 M sodium phosphate buffer, pH 7.2, and dialyzed against 50 vol oft.his buffer. The dialysate (14 ml, 89 Azs0/ml ) was applied to a DEAE-Sephadex column of approximately 50 ml bed volume equilibrated with the above buffer. The column was washed with 50 ml of buffer (above) then eluted with 50 ml of buffer containing 1.0 M NaC1. Column fractions were assayed by inhibition of poly(C)-dependent ATPase as previously described (Kent and Guterman 1981).

Results

BLA

1

2

3

4

Fig. 1. The effect of the presence of antibody to Rho factor during in vitro synthesis of Rho factor. The samples were boiled for 2 rain before application to the gel. The DNA templates used were from the rho + plasmid pJG32 (lanes 3 and 4) and the lacZ + 2h8Odlac phage (lanes 1 and 2). Antibody to Rho factor (1 gl) was present for the samples shown in lanes 1 and 3, but was not present in the parallel samples in lanes 2 and 4

Rho Factor Synthesis in vitro is Stimulated by Antibody to Rho Factor

Plasmid pJG32 contains the rho gene on an 8 kilobase (kb) HindIII segment cloned in plasmid pBR322 (Gray et al. 1981). In vitro this plasmid directs the synthesis of the 50,000 dalton Rho factor and lesser amounts of the vector coded bla gene product (Fig. 1, lanes :3, 4 and Fig. 2, lanes 14). This is the expected result based on the proteins synthesized in maxicells by this plasmid (Gray et al. 1981) and by plasmid pMCS1 derivatives that carry the 8 kb HindIII segment (Calhoun et al., accompanying article). Antibody to Rho factor stimulates Rho synthesis in vitro (compare Fig. 1, lanes 3 and 4, Fig. 2, lanes 3 and 4 and Fig. 4, lanes 4 and 5). The identity of the 50,000 dalton protein produced in vitro as Rho factor is confirmed by the observation that in the presence of specific antibody it is precipitated or forms an antigen-antibody complex that does not enter the SDS gel (Fig. 2, lane 3). The effect of antibody to Rho factor was tested using a variety of phage and plasmid D N A templates (see Discussion). For example, Fig. 1, lanes I (aatibody present) and 2 (antibody absent) reveal that several 2h8Odlac coded proteins are unaffected or slightly inhibited (e.g.,/%galactosidase) by antibody to Rho factor. The stimulatory effect of antibody to Rho factor during in vitro synthesis is, therefore, specific for the rho gene, at least for the concentrations used. Rho Factor Synthesis in vitro is Inhibited by the Addition o f Rho Factor

Rho factor synthesis is inhibited when purified Rho protein is added to the in vitro system (Fig. 3, lane 2), compared

I

2

3

4 TOP

RHO

BLA

Fig. 2. Specificity of the effect of antibody to Rho factor during in vitro synthesis of Rho factor. The samples were not boiled prior to application to gel and the antigen-antibody complex is retarded at the top of the gel (lane 3). The DNA template was the rho + plasmid pJG32. Antibody to Rho factor was present in lane 3 but not in lane 4. As a control, antibody to two other purified E. coli K-12 proteins were present for the samples shown in lane 1 (antibody to transaminase B) and lane 2 (antibody to threonine deaminase)

212

RHO

to the control (Fig. 3, lane 5). Addition of the mutationally altered rho-ll5 gene product does not inhibit Rho synthesis (Fig. 3, lane 1). This mutant R h o protein is overproduced in vivo (Ratner 1976), an observation that is consistent with the in vitro result. The inhibitory effect of R h o factor is specific for the rho gene at the concentrations used, but a nonspecific inhibition of many genes is seen with higher levels of Rho factor (unpublished observations).

Antibody Reverses the Effect of Rho Factor During in vitro Synthesis BLA

1

2

3

4

Discussion

5

Fig. 3. The effect of the presence of Rho factor during in vitro synthesis of Rho factor. The in vitro synthesis of Rho factor from plasmid pJG32 was examined in the absence (lane 5) or presence (lane 2) of 0.5 gg of wild type Rho factor. The addition of 0.5 gg of Rho factor from the rho-ll5 mutant (lane 1) or two unrelated purified proteins (lane 3, 2 I~g of transaminase B; lane 4, 1.5 I~g of threonine deaminase) were without effect

RHO

BLA

I

2

3

4

The inhibition of Rho factor synthesis in vitro by added Rho factor (Fig. 4, lane 3) is reversed by the simultaneous addition of antibody to Rho protein (Fig. 4, lanes I and 2). These results demonstrate the specificity of the effects in vitro of Rho factor and antibody to Rho factor.

5

Fig. 4. Reversal of the effect of antibody to Rho factor by purified Rho factor. The samples (3 gl) were boiled before application to the gels. The in vitro synthesis of Rho factor from plasmid pJG32 was examined in a control to which no additions were made to the reaction mixture (lane 5), or samples containing 0.3 ~tl of antibody to Rho (lane 4), 0.5 gg of wild type Rho factor (lanes 1-3), and antibody to Rho (0.3 I11 in lane 2 and 1.0 gl in lane 1). The gel was overexposed to the film to permit clear visualization of the greatly reduced Rho protein in lane 3. The stimulatory effect of antibody alone (lane 4) compared to the control (lane 5) is more apparent on shorter exposures of the gel to film that approximate a linear dosage response region

The results presented here substantiate earlier proposals based on indirect evidence that the expression of the rho gene is controlled by its gene product, the Rho protein. Following the discovery of the first known examples of autoregulation (reviewed by Calhoun and Hatfield 1975), additional genes with this pattern of regulation have been described. Autoregulation is seen with genetic regulatory proteins [e.g., the phage lambda cI repressor (Ptashne et al. 1980), the trpR repressor (Kelley and Yanofsky 1982), and the araC activator protein (Casadaban 1976)], with structural proteins [e.g., the ribosomal genes (reviewed by Lindahl and Zengel 1982)] and with a protein such as the phage T4 gene 32 protein that participates in D N A recombination, replication, and repair (discussed by Lemaire et al. 1978). Autoregulation has been reported to occur at the level of transcription (e.g., phage lambda cI repressor) and translation (e.g., ribosomal genes). It is not known if the autoregulation seen for the rho gene occurs at the level of transcription or translation, and it is not known if other metabolites or proteins affect this activity. It is particularly striking that the purified mutant Rho factor coded by the rho-115 gene is deficient in genetic regulatory activity in vitro (Fig. 3, lane 1). This result is consistent with the observation of Ratner (1976) that the rho-ll5 mutant overproduces the altered protein in vivo. Consideration of the altered properties of the well characterized rho-115 gene product (Kent and Guterman 1978) might be expected to give some insight into the possible mechanism of autoregulation of gene expression by R h o factor. The rho-ll5 coded protein appears to be defective in its interactions with both R N A polymerase and R N A . It is not possible to infer, therefore, whether the normal genetic control by R h o is mediated at the transcriptional or translational level. In addition, there is no obvious correlation between increased Rho factor synthesis in vivo and the properties of Rho coded by various mutant alleles. Brown et al. (1982) recently reported evidence for attenuator control of the rho gene. Rho factor is thought to affect many genes during normal termination and during specific regulatory events (see Introduction). The effect in vitro of the addition of purified Rho factor or antibody to Rho factor might be expected to vary in a gene specific manner. Indeed, under the same in vitro experimental conditions used here, antibody to Rho

213 specifically increases the expression of some genes and decreases the expression of others, eve~L when the genes are tested together on the same plasmid D N A template (unpublished observations). I n addition, we have observed effects of the presence of Rho factor purified from the m u t a n t and from the wild type using other D N A templates. Their effects on some genes are similar, in contrast to their dramatic differences seen using the rho gene itself (Fig. 2). Thus, the residual activity of the r h o - l l 5 gene product is selective in that only some target sites are affected. Further studies will be required to elucidate the specific mechanism of the action in vitro of Rho factor and its antibody on rho and other genes. Acknowledgements. These experiments were supported by NIH grants GM-23182 (DHC) and GMS-24461 (SKG). DHC is the recipient of an Irma T. Hirschl Career Scientist Award.

References Adhya S, Gottesman M, DeCrombrugghe B (1974) Release of polarity in Escherichia coli by gene N of phage lambda: Termination of transcription. Proc Natl Acad Sci USA 71:2534-2538 Brown S, Albrechtsen B, Petersen S, Klemrn K (1982) Localization and regulation of the structural gene for transcription-termination factor rho of Eseherichia coli. J Mol Biol 162:283-298 Calhoun DH, Hatfield GW (1975) Autoregulation of gene expression. Annu Rev Microbiol 29:275-299 Casadaban M (1976) Regulation of the regulatory gene for the arabinose pathway, araC. J Mol Biol 104:556-557 Gray JE, Patin DW, Calhoun DH (198i) Identification of the protein products of the rrnC, ilv, rho region of the Escherichia coli K-12 chromosome. Mol Gen Genet 183:428-436 Greenblatt J, McLimont M, Hanly S (1981) Termination of transcription by NusA gene protein of £'scherichia coli. Nature 292 :215-220 Guterman SK, Howitt CL (1979) Rifampicin supersensitivity of rho strains of E. coli, and suppression by sur mutation. Mol Gen Genet 169:27-34 Housley PR, Leavitt AD, Whitfield HJ (1981) Genetic analysis of a temperature sensitive Salmonella typhimurium rho mutant

with an altered Rho associated polytidylate dependent adenosine triphosphate activity. J Bacteriol 147:13-24 Imai M, Shigesada K (1978) Studies on the altered Rho factor in nitA mutants of Escherichia coli defective in transcription termination. I. Characterization and quantitative determination of Rho in cell extracts. J Mol Biol 120:451466 Kelley RL, Yanofsky C (1982) trp aporepressor production is controlled by autogenous regulation and inefficient translation. Proc Natl Acad Sci USA 79:3120-3124 Kent RB, Guterman SK (1981) A mutant Rho ATPase from Escherichia coli that is temperature sensitive in the presence of RNA. Mol Gen Genet 181:367-372 Kung H-F, Morrissey J, Revel M, Spears C, Weissbach H (1975) Studies on the lactose operon. The control of DNA directed in vitro protein synthesis by interferon factor i-a. J Biol Chem 250: 8780-8784 Lemaire G, Gold L, Yarus M (1978) Autogenous translational repression of bacteriophage T4 gene 32 expression in vitro. J Mol Biol 126:73-90 Lindahl L, Zengel JM (1982) Expression of ribosomal genes in bacteria. Adv Genet 21:53-121 Platt T (1981) Termination of transcription and its regulation in the tryptophan operon of E. coli. Cell 24:10-23 Ptashne M, Jeffrey A, Johnson AD, Maurer R, Meyer BJ, Pabo CO, Roberts TM, Sauer RT (1980) How the lambda repressor and cro work. Cell 19:1-11 Ratner D (1976) Evidence that mutations in the suA polarity suppressing gene directly affect termination factor Rho. Nature 259:151-153 Reyes O, Gottesman M, Adhya S (1976) Suppression of polarity of insertion mutations in the gal operon and N mutations in bacteriophage lambda. J Bacteriol 126:1108-1112 Richardson JP, Grimley C, Lowry C (1975) Transcription termination factor Rho activity is altered in Escherichia eoli with suA gene mutations. Proc Natl Acad Sci USA 72:1725-1728 Roberts JW (1969) Termination factor for RNA synthesis. Nature 224:1168-1174 Yanofsky C (1981) Attenuation in the control of expression of bacterial operons. Nature 289:751-758 Communicated by G.A. O ' D o n o v a n Received July 16, 1983