Proc. Nat. Acad. Sci. USA Vol. 72, No. 5, pp. 1725-1728, May 1975

Transcription Rho Activity Is Altered in with suA Gene Mutations (RNA-dependent ATPase/RNA polymerase/polarity) JOHN P. RICHARDSON, CHRISTOPHER GRIMLEY, AND CAROLYN LOWERY Department of Chemistry, Indiana University, Bloomington, Ind. 47401 Communicated by James D. Watson, February 51, 1975

ABSTRACT has been purified from a strain allele for the suA gene have been purified to of E. coli containing the Su78 mutation in the suA gene homogeneity. and assayed in another strain with an amber mutation in The yields in both cases are the same: about 0.12 mg of pro- the suA gene. The rho from the Su78 mutant strain is tein from 20 g of cells, which is similar to yields from other present in normal amounts but has altered termination cells by means of the Roberts purification procedure. The two function; it does not terminate at some sites pure have the same specific activities for the poly(C)- that are recognized effectively by the rho factor from the activated of isogenic wild-type strain. Rho in cells with an amber hydrolysis ATP-20 ,umol of ATP hydrolyzed mutation in the suA gene has been assayed by its RNA- min' per mg-and the same mobilities when electrophoresed dependent ATPase activity. Extracts of cells of this strain on polyacrylamide gels containing sodium dodecyl sulfate (Fig. have only 9% as much of this rho activity as extracts of 1). However, these rho factors differ significantly in their abili- cells of the isogenic wild-type strain. These results suggest ties to depress the yields of RNA synthesized in vitro with that rho is the product of the suA gene. Since mutations in the su,4 gene are known to decrease polar effects of purified RNA polymerases from several DNA templates mutations in other genes, it is also suggested that rho (Table 1). For each DNA, the rho from the 8uA mutant factor is at least partially responsible for polar effects. strain (Su78) reduces the yield of RNA synthesized less than the rho from the cells with the wild-type allele. The magnitude Many nonsense and frameshift mutations in one gene of an of this difference depends on the DNA used. With T4 DNA, pleiotropically reduce the level of expression of the the Su78 rho caused only an 8% depression, compared with genes in the operon that lie on the operator-distal side of the the early 50% depression with the wild-type rho. In contrast, mutant gene (1). This polar effect is evident in the levels of with PM2 DNA the Su78 rho caused a 33% depression, both the messenger RNA sequences corresponding to the whereas the wild-type rho still only caused a 50% depression. affected genes and the products (2, 3). In Escherichia These differences are not because of limitations in the amount coli, second-site mutations in the suA gene can partially re- of rho used; the reactions were saturated with the rho used in lieve the polar effect of the original mutation without sup- each case. pressing the orignal mutation (4, 5). Since these suA gene Analyses of the effects of these factors on the size of the mutations are recessive to the wild-type allele (5) and since RNA made on T4 and T7 DNA templates confirm the inter- amber (nonsense) 8uA mutations have been found (6), it has pretation that the depression of yields of RNA synthesis given been suggested that the product of the wild-type suA gene is a in Table 1 is an accurate indication of the relative termination protein required for the full polar effect (6). In this paper we activities of the factors. With T7 DNA, for example, the present evidence that the product of the suA gene is the rho sedimentation profile of the RNA synthesized in the absence transcription termination factor. of rho has a peak indicating a sedimentation coefficient of 17.8 Rho factor was first isolated and purified from E. coli K12 S in 1.1 M formaldehyde (Fig. 2), which corresponds to an by Roberts, who showed that it causes specific termination of RNA with a molecular weight of 2.2 X 106 (11). When the rho the synthesis of X RNA molecules in vitro (7). It has since been from the wild-type cells is present, the RNA peak is at 9.8 S, shown that rho is active in terminating transcription from a which corresponds to an RNA with a molecular weight of 0.5 large number of natural DNA templates (8). Roberts also X 106. However, when rho from Su78 is present, most of the showed that rho is not a ribonuclease; it does not cleave or RNA molecules synthesized are much larger than those syn- degrade large, isolated X RNA molecules even in complete thesized in the presence of the wild-type rho; the main peak is RNA polymerase reaction mixtures (7). We now know that it at 16.8 but there are also at 14.5 S and does S, peaks 12.0 S. Thus the not catalyze the degradation of nascent RNA molecules Su78 rho does not terminate T7 transcription at the site recog- either (9). On the other hand, when RNA molecules are pres- nized by the wild-type rho it does ent, factor, although appear to rho does catalyze the hydrolysis of nucleoside triphos- terminate the transcription of some RNA molecules at phates to nucleoside diphosphates and orthophosphate (9). other Although the sites that are reached later by RNA polymerase. significance of this ATPase activity is not yet Although the two rho factors are identical with clear, it can be used respect to for a convenient quantitative assay of rho their poly(C)-activated ATP hydrolysis activities, they do factor. We have used this assay to identify and purify rho show factor from cells different rates of ATP hydrolysis when activated by with mutations in the suA gene. nascent RNA molecules. In complete RNA polymerase reac- RESULTS tion mixtures containing T4 DNA or T7 DNA, only the rho that causes extensive termination also catalyzes the release of The rho proteins from a 8uA - strain isolated by Carter and Pi from [y-32PJATP (Table 2). This observation is consistent Newton (Su78) (10) and an isogenic strain with the wild-type with others we have made (C. L. and J.P.R., unpublished ex- 1725 Downloaded by guest on September 25, 2021 1726 Biochemistry: Richardson et al. Proc. Nat. Acad. Sci. USA 72 (1975) TABLE 1. Termination activity of rho factors from cells with the Su78 mutation and cells with the wild-type allele in the suA gene

[3H]UMP Phage incorporated Inhibition DNA Rho (cpm) (%) T4 727 Wild-type 396 46 Su78 670 8 T5 - 2136 Wild-type 1210 43 Su78 1919 10 T7 - 1649 Wild-type 660 60 Su78 1399 15 fd RFI - 1934 Wild-type 554 71 Su78 1064 45 PM2 - 1715 Wild-type 861 50 Su78 1151 33

Each reaction mixture contained 0.04 M Tris HCl (pH 7.9); 0.05 M KCI; 12 mM MgCl2; 0.1 mM EDTA; 0.1 mM dithio- threitol; 0.7 mg/ml of bovine serum albumin; 0.2 mM each of ATP, GTP, and UTP; 0.12 mM [3H]CTP (10 Ci/mol); 0.3 pg of RNA polymerase; 0.8 ug of DNA; 0.08 ug of rho factor, where A B C appropriate, in a final volume of 0.05 ml. After incubation for FIG. 1. Polyacrylamide gel electrophoresis of purified rho 30 min at 370, the reaction was stopped by the addition of 0.4 ml proteins. Rho protein was isolated by the procedure of Roberts of 0.1 M Na4P207 and 0.5 ml of 10% trichloroacetic acid. Precipi- (7) from E. coli strains T82 (Su78) and T83 (suA +) supplied by tated RNA was collected and washed on membrane filters and the Austin Newton. These strains are isogenic P1 transductants of a radioactivity on the dried filters was measured in a toluene-base ilv- derivative of MU118 (ZU118) (13). The original Su78 strain scintillation fluid. The procedures for isolation of RNA polymer- was described by Carter and Newton (10). The proteins were ase (from E. coli B) and have been described previously electrophoresed in a discontinuous slab gel (22), using a 3.5% (24, 25). Wild-type and Su78 rho are the purified factors de- stacking gel and a 10% running gel of polyacrylamide in the scribed in Fig. 1. sodium dodecyl sulfate/Tris. HC1 buffer system described by Laemmli (23). The proteins were stained with Coomassie brilliant activity for the rho isolated from 20 g of cells with a suA amber blue using the conditions of Laemmli (23); (A) 0.2 pg of rho from mutation (strain 2055) as was found for the rho isolated from T83 (wild-type); (B) 0.1 pug each of rho from T83 and T82 (Su78); 20 g of cells of an isogenic strain (2034) with the wild-type (C) 0.2 pg of rho from T82. allele in the suA gene. In both cases, the rho eluted from the phosphocellulose in a single sharp peak in 0.16 M potassium periments) correlating rho termination activity with its na- phosphate buffer. . scent-RNA-dependent ATP hydrolysis activity. For instance, Since the residual level of rho factor in strain 2055 is higher in 0.12 M KCl, the rho from E. coli B no longer terminates the than would be expected for the product of a gene with an transcription of T7 DNA, and the T7 RNA made under these amber mutation, caution must be taken in concluding that rho conditions also does not activate the rho ATP hydrolysis is the product of the suA gene. Mutations in that gene could reaction. Yet under these same ionic conditions, the rate of affect rho indirectly; the suA gene product might be an en- ATP hydrolysis in the poly(C)-activated reaction is the same zyme that modifies or activates rho. Nonetheless, whether the as in 0.05 M KCl. The requirements for activation of the effect is direct or indirect, one conclusion is clear: rho factor ATPase activity of rho by nascent RNA molecules are ap- termination activity is reduced in the polarity suppressor parently more stringent than the requirements for activation strains. This implies that rho is involved in translational by poly(C). polarity. Further evidence that the product of the suA gene is rho DISCUSSION has come from attempts to isolate rho from one of the strains isolated by Morse and Guertin (6) containing an amber muta- From our studies of the mechanism of rho action we propose tion in the suA gene. Although it is not possible to make re- the following model for how rho causes polar effects. We sug- liable quantitative estimates of the amount of rho present in gest that termination mediated by rho factor depends on the crude extracts of cells with any of the current assays, rho recognition by rho of a particular sequence or structure of the factor can be unambiguously identified and quantitatively nascent RNA as it emerges from RNA polymerase. If a ribo- assayed after chromatography on phosphocellulose. At this some is able to translate that RNA as it emerges from the stage, there was only 9% as much poly(C)-activated ATPase RNA polymerase, it could interfere with this recognition pro- Downloaded by guest on September 25, 2021 Proc. Nat. Acad. Sci. USA 72 (197-6) Rho from suA Mutants 1727

TABLE 2. Correlation of rho termination activity with rho Sedimentation coefficient (S) ATPawe activity 20 15 10 5 0

3P released as PPi Pi DNA Rho (pmol) (pmol) x E T4 - 140 6 C) Wild-type 58 85 Su78 165 9 z T7 - 199 <1 I-F Wild-type 67 68 Su78 196 11

The reaction conditions were identical to those of Table 1 2 3 4 5 except that [yy-32P]ATP (3800 cpm/nmol) was used instead of Volume of tube from bottom (ml) [3H]CTP and the amounts of rho were 0.4 yg per mixture. After FIG. 2. Sedimentation distribution of T7 RNA synthesized in incubation for 30 min at 370, the amounts of 32P released as Pi 0.05 M KCl in the presence of wild-type or Su78 rho or in the and PPi were analyzed as described previously (9). The release of absence of rho. Each reaction mixture contained 0.04 M Tris- HCO Pi is corrected for 11 pmol released in the absence of any enzyme (pH 8.0); 0.05 M KCl; 12 mM MgC12; 0.1 mM EDTA; 0.1 mM components. dithiothreitol; 1.6 mM each ATP, GTP, and UTP and 1.0 mM [3H]CTP (50 Ci/mol); 0.8 Izg of RNA polymerase; 2 ,Ag of T7 cess, thus preventing rho-mediated termination. However, DNA, and 0.2 ,ug of rho protein, where appropriate, in a final when translation is prematurely terminated at a nonsense volume of 0.05 ml. After incubation for 30 min at 370, a 30 ,ul codon, the would leave the RNA and expose a seg- sample was removed, treated with detergent, EDTA, and formal- ment of the emerging nascent RNA to rho action. dehyde as described previously (24), and the treated sample was The involvement of rho in translational polarity would also layered on a 5 ml 20-5% linear sucrose gradient containing 0.1 M sodium phosphate buffer (pH 7.7) and 1.1 M formaldehyde. depend on the presence of rho recognition sites within cistrons. Centrifugation was for 15 hr at 29,000 rpm in a Spinco SW 50.1 Studies of transcription of the in vitro have indi- rotor at 4°. Sedimentation coefficients were estimated from the cated that there is a site recognized by rho in the middle of the peaks of the marker E. coli total RNA (11). 0, without rho; z gene, the gene for fl-galactosidase (12). One striking char- 0, with wild-type rho; A, with Su78 rho. acteristic of polar mutants in the z gene is that the severity of the polar effect is dependent on the position of the mutation clease activity that is present in extracts of wild-type cells. in the gene; amber mutants near the beginning of the z gene Although pure rho is not itself a ribonuclease, it could possibly give a much stronger polar effect than those near the end of be a component of the endonuclease responsible for the activ- the gene (13). Similar gradients of polar responses have also ity detected by Kuwano et al. Thus, we cannot conclude cate- been found for mutations in genes for the tryptophan operon gorically that rho's function in causing polarity is not due to of E. coli (14). A gradient of polarity could be explained by rho degradation of nascent RNAs exposed when dis- action, if there are several rho termination sites within a gene sociate at a nonsense codon. However, since it is known defi- and if the sites are not all totally effective. In this case there nitely that rho does terminate transcription in vitro, we prefer would be more chances for rho to terminate transcription when the hypothesis involving this mechanism for its in vivo func- a ribosome leaves the nascent RNA early in the gene than tion as well. This hypothesis is also consistent with the ob- when it leaves it late in the gene. servations of Imamoto (21) that trp RNA molecules are not It is not known yet whether there are several rho sites synthesized in vivo beyond nonsense mutations in the trp within the z gene, but it is a large gene with at least 3600 base operon. In order to remove the uncertainty that remains con- pairs (3) and rho sites could be distributed as frequently as one cerning the involvement of rho factor in polarity, it will be for every 800 to 1000 nucleotides transcribed (15). The effi- necessary to analyze further rho function in cellular metabo- ciency of rho action in the cell is also not known, but it would lism. depend on the amount present, and the current estimate, which could be low by a factor of one hundred, indicates that We thank Drs. A. Newton and D. E. Morse for kindly supply- there is only one rho molecule for 100 RNA polymerase mole- ing the bacterial strains; Dr. T. Blumenthal for a timely sugges- cules In contrast, completely effective rho termination tion; Dr. B. Alberts for encouraging this investigation; and (16). Helene Cann for technical help with the sedimentation experi- in vitro requires one rho per RNA polymerase (17). ments. This investigation was supported by Research Grant Since rho must also interact with RNA polymerase to cause AI-10142 from the National Institute of Allergy and Infectious termination of RNA synthesis, an alteration of the enzyme Diseases. J.P.R. is the recipient of Research Career Development could also affect rho action. This could be the way N gene Award GM-70422 from the National Institute of General Medical product in A acts to cause transcription to continue in vivo past Sciences. sites where rho is known to function in vitro (7). This same 1. Zipser, D. (1969) Nature 221, 21-25. alteration could also overcome the polar effects of nonsense 2. Imamoto, F. & Yanofsky, C. (1967) J. Mol. Biol. 28, 1-23. mutations in bacterial genes transcribed by read-through from 3. Contesse, G., Crepin, M. & Gros, F. (9170) in The Lactose Operon, eds. Beckwith, J. R. & Zipser, D. (Cold Spring X genes (18, 19). Harbor Laboratory, Cold Spring Harbor, N.Y.), pp. 111- Kuwano et al. (20) have presented evidence that extracts of 141. cells containing mutations in the suA gene lack an endonu- 4. Beckwith, J. (1963) Biochim. Biophys. Acta 76, 162-164. Downloaded by guest on September 25, 2021 1728 Biochemistry: Richardson et al. Proc. Nat. Acad. Sci. USA 72 (1976)

5. Morse, D. E. & Primakoff, P. (1970) Nature 226, 28-31. 15. Takanami, M., Okamoto, T. & Sugiura, M. (1971) J. Mol. 6. Morse, D. E. & Guertin, M. (1972) J. Mol. Biol. 63, 605-608. Biol. 62, 82-88. 7. Roberts, J. W. (1969) Nature 224, 1168-1175. 16. Minkley, E. G., Jr. (1973) Ph.D. Dissertation, Harvard 8. Goldberg, A. R. & Hurwitz, J. (1972) J. Biol. Chem. 247, University. 5637-5645. 17. Richardson, J. P. (1970) Cold Spring Harbor Symp. Quant. Biol. 35, 127-133. 9. Lowery-Goldhammer, C. & Richardson, J. P. (1974) Proc. 18. Adhya, S., Gottesman, M. & de Crombrugghe, B. (1974) Nat. Acad. Sci. USA 71, 2003-2007. Proc. Nat. Acad. Sci. USA 71, 2534-2538. 10. Carter, T. & Newton, A. (1971) Proc. Nat. Acad. Sci. USA 19. Franklin, N. C. (1974) J. Mol. Biol. 89, 33-48. 68, 2962-2966. 20. Kuwano, M., Schlessinger, D. & Morse, D. E. (1971) 11. Boedtker, H. (1968) J. Mol. Biol. 68, 61-70. Nature New Biol. 231, 214-217. 12. de Crombrugghe, B., Adhya, S., Gottesman, M. & Pastan, I. 21. Imamoto, F. (1970) Nature 228, 232-235. (1973) Nature New Biol. 241, 260-264. 22. Studier, F. W. (1973) J. Mol. Biol. 79, 237-248. 13. Newton, A., Beckwith, J. R., Zipser, D. & Brenner, S. 23. Laemmli, U. K. (1970) Nature 227, 680-685. (1965) J. Mol. Biol. 14, 290-295. 24. Richardson, J. P. (1973) J. Mol. Biol. 78, 703-714. 14. Yanofsky, C., Horn V., Bonner, M. & Stasiowski, S. (1971) 25. Richardson, J. P. & Parker, S. R. (1973) J. Mol. Biol. 78, Genetics 69, 409-433. 715-720. Downloaded by guest on September 25, 2021