576 Anderson during this work. We also wish to thank Mr. R. W. Kent and Mr. D. Smith for their help in connexion with the electrostatic procipitator, and ‘Mrs. A. Porry for assistance in the chemical analyses. ‘Steyart, NG, Osmond, R. 0. D., Grogks, B. N., and Fisher, EAM, Te Xtomio erty ‘osedren”Betablihmint, “Harwell, P| 264 G08) 1, Burton, 1. K., Crookal, J. 0., and 166, 985 (1060) 1D. HL, Crooks, R, N. and Fisher, Hat B., “Hovedreh Hitablihment, Harwell, 3868 (1000). Martel, B. A., Science, 128, 1197 (05 Brewer, A. W., Quart. J. Roy. Met. So. 78, 361 (1040). Dobwon, @.M.'B., Proc. oy. Soe, A, 286, 157 (1050). Feely, H. W., and Spa, J., ature, 28, 1082 (1060), * Anderson, W., Bentey, “Genta, 0a Wa Atomic Boergy NATURE May 13, 1961 vou. 190 “ange. Bate. aro, eK nd ere ©. ‘Nature, 186, 223 (1980). + anda, Wh penkay SM. Parts, RP. rohal 3.0, and ‘Burton, L. K., Nature, 187, 650 (1960). ss nprnetedh, Be arte Poe 18 S90 Go), nERIge Wen 6.8 Conroe error Pala fom Mater ‘Weapons Testa, 8, 2229 (1950). > pata Mom Sepa seme, Rp. DAS, Watntn ve nat DF tnd aerate Sh 06, ttn Ne Ompts Ops Ont amd Wes 8 ML, Pog eg pee © wan ere 8 0 0 one A oman Bese Faber, ‘and ‘Bvett, T. W., Atomic iinergy Research Sls lied! 10k Danced Belg on Fatt fr Bucher Woops Slt Ee tl AN UNSTABLE INTERMEDIATE CARRYING INFORMATION FROM GENES TO RIBOSOMES FOR PROTEIN SYNTHESIS By Dr. S. BRENNER Medical Research Council Unit for Molecular Biology, Cavendish Laboratory, University of Cambridge DR. F, JACOB Institut Pasteur, Paris AND Dr, M. MESELSON Gates and Crellin Laboratories of Chemistry, California Institute of Technology, Pasadena, California A. LARGE amount of evidence suggests that gonetio information for protein structure is encoded in deoxyribonucleic acid (DNA) while the actual ‘assombling of amino-acids into proteins occurs. in eytoplasmic ribonucleoprotein particles called ribo. somes. ‘The fact that proteins are not synthesized direotly on genes demands the existence of an intermediate information carrier. ‘This intermediate template is generally escumed to be a stable ribo- nucleic acid (RNA) and moro specifically the RNA of the ribosomes. According to the present. view, each geno controls tho synthesis of one kind of specialized ribosome, which in tum directs. the synthesis of the corresponding protein—a scheme which could be epitomized as the one geno-ono ribosome-one protein hypothesis. In tho past fow years, however, this model has ‘ericountered some Gifficulties: (1) ‘The remarkable homogeneity in size and nucleotide composition? of the ribosomal RNA reflects neither the range of size of palypeptic ‘chains nor the variation in the nucleotide composition observed in the DNA of difforent: bactorial species*., (2) The capacity of bacteria to synthesize a given protein does not seem to survive beyond the integrity. of the ing geno’. (3) Regulation of protein synthesis in bacteria seems to operate at the lovel of the synthesis of the information intermediate by tho geno rather than at the level of the synthesis ‘of tho protein’. ‘Thoso results aro compatible with the existence of stable RNA intermediates acting as templates for protein synthesis. The paradox, however, can be resolved by the hypothesis, put forward by Jacob and Monod, that the ribosomal RNA is not the intermediate oarvir of information from. to protein, but ratl ribosomes. fre non-apecalied structures which reesive gonctie information from the geno in the form of an unstable intermodiato or ‘messenger’. We present hero tho results of experiments on phage-infected bacteria which give direct support to this hypo. thesis. wat ‘ ‘When growing bacteria aro infected with a virulent bacteriophage such as 72, synthosis of DNA stops immediately, to resume 7'min. later‘, while protein synthesis continues ata constant ‘Tate’, After infection many bacterial enzymes aro no longer produced"; in all likelihood, the new protein is genetically determined by the phage. A large number of new enzymatic activities appears in the infected cell during the first fow minutes following infection’, and from the tenth minute onwards some 60 per cent of the protein synthesized can be accounted for by the proteins of the phage coat?. Surprisingly enough, protein synthesis after infection is not accompanied, as in growing cells, by a not synthesis of RNA‘. Using isotopic labelling, however, Volkin and Astrachan™ were able to demonstrate high turnover in a minor RNA fraction after infection. Most remarkable is the fact that this RNA fraction has an apparent nucleotide ‘composition which to that of the DNA ‘of the phage and is markedly different from that of the host RNA™, Recently, it has been shown ‘that the bulk of this RNA is associated with the ribosomes of the infected cell’ Phage-infeoted bacteria therefore provide a situ. ation in which the synthesis of protein is suddenly veheno.4776 May 13, 1961 NATURE 877 Model UninFected Celt Infected Cell Hoot Tone inhivited 1 ceoo Looe] maton sees 3 3 PRoTEW Phage ONA a DNA ; = Ae Fig. 1. switched from bacterial to phage control and without the concomitant synthesis of stable RNA. 4 priori, threo typos of hypothesis may be considered to account for the known facts of phage protein synthesis (Fig. 1). Model I is the classical Model. After infeotion the bacterial machinery is switched off, and new ribosomes are then synthesized by the phage genes. Tho ad hoe hypothesis has to be added that these ribosomes aro unstable, to account for the tumover: of RNA after phage infection. 'This is, in fact, the model favoured by Nomura etal.%, Model IT assumes that in the particu- lar case of phage the protoins aro aasernbled directly on the DNA; the now RNA is a special molecule which enters old ribosomes and destroys their capacity for protein synthesis. At the same time, synthesis of ribosomos is switched off. Model II implies that a special type of RNA molecule, or ‘messenger RNA’, exists which brings genetic infor- mation from genes to non-specialized ribosomes and that tho consequences of phage infection are two-fold: (a) to switch off. the synthosis of new ribosomes ; (0) to substitute phage messenger RNA for bacterial messenger RNA. This substitution oan occur quickly only if mesenger RNA is unstable ; ed ; 3 ae peas. ‘Three models of information tranafer in phage-infected cells the RNA made after phago infection doos turn over and appears, therefore, as 8 good oandidato for tho mossonger. ‘Tt is possible to distinguish experimentally between. these three models in the following way: Bacteria fare grown in heavy isotopes 60 that all cell oon- stituents aro uniformly labelled ‘heavy’. ‘They are infected with phage and transferred immediately to @ medium containing light isotopes so that all constituents synthesized after infection are ‘light’. ‘Tho distribution of new RNA and now protein, labelled with radioactive isotopes, is then followed by donsity gradient contrifugation'® of purified ribosomes. Density gradient centrifugation was carried out ina tive centrifuge, and tho ribosomes were stabilized by including magnesium acetate (0-01- 0:06 Mf) in the emsium chloride solution. Ribosomes show two bands, a heavier A band and a lighter B band, tho relative proportions of which, for ‘@ given proparation, dependvon the magnesium concentration used. | The lower the magnesium ‘eoncentration, the amaller the proportion of B band ribosomes and the larger the proportion of A band ribosomes. 23318 NATURE Fu pores 800 o8 0 po 00 oat B00 | os! au 300 02 200 oa “5 wo 20 $0 40 50 0 70 80 00 100 Fraction No. 9,2: , Digtlbutlon of hoavy and light ribosomes in a donsity ersten m sim. of saogiam contarning Fn ng 00 a8 200 or 700 a8 000 05 4 5 so0 z os 4008 08 so0 oe 200 o1 100 Fr a Fraction No. Fu; 3,_Dielbution of randomized heyy and ght ome, Ins ena aratent. othe mire of CAO ang MAGNE “iaivacd ay for 1 he. against, 0-0005 {pS hcephate bar 7H 70 an then tee 0 hrs againat two changes Tnguclun ace O'o0r aftr buer pH T-41 sagm of rhowomes Yas ceatefuged for heat Si. i cena chloro containing 0 03 ‘magnesium acetate,” ‘Fhe drops were assayed or ultra-violet ‘sbsorption (0) ahd "P eontent (@) In order to show that there is no aggregation of ribosomes during preparation and density gradient centrifugation an experiment was carried out on ribosomes extracted from mixture of *N0 and HNUQ bacteria. The results are shown in Fig. 2, from which it can be seen that ribosomes of different isotopic compositions band independently and that there aro no intermediate classes. The same pro- eration was then dialysed against low magnesium 24 May 13, 1961 vor. 190 to dissociate the ribosomes into their 50 S and 30S ‘components and thon against high magnesium to re-associate thé sub-units, This should have resulted in distributing heavy 308 and 50 S sub- units into mixed 70 Sand 100 S ribosomes. Sur- Disingly enough, ‘density gradiont oontrifugation Preparation (Fig. 3) yiolds tho samo bands as found i tho eigial sooosee ‘opt fora doeroate fa te bropactin Of Hin eo means that both bands contain units which do not undergo reversible association and dissociation and that the mixed 70 § ribosomes prepared by dialysis separate Bus or os os on os on on a rr ee ey Fraction No. Pig 4 a 00 os oF os Countafmtn on os Cy 02 © oa 0 102 «9 40 50 «10 Fraction“No, Wig. 5 Pigs. 4 and 5. Distsibution and turnover of HA formed after altar of ci Be natant reglrng ine and urael) was Tafeetod with 74D (oultplety 30) ‘Reach 1 ea) fom ra 9 Aft fn fer nection, ince appresinaily 3g. of Pala opines ‘ppsinsicy 3 mgm of Pr emines fot te hr at 37.000 r'pam. i emaiom chords A nagnesitm tovtute? “Aiernate trope were in ur forty ile sarin (0) fon pe en trchormcetie a i an ate, ad he ‘on ambrans See Ser Sees tat Sie Sa ars etal oes Sir aricbetan (8) whe