Perspectives

Anecdotal, Historical And Critical Commentaries on Genetics Edited by James F. Crow and William F. Dove

The Amber Mutants of Phage T4

Franklin W. Stahl

Institute of Molecular Biology, Un.iversity of Oregon, Eugene, Oregon 97403-1229

A MAJOR effort of today’s biology is the analysis of with the related phage T4 and showed that the high- development by genetic methods,an approach so dose slope of multicomplexes was 0.4 of that for mono- successful that previously unimaginable insights have complexes. become commonplace. Amajor moment in thegrowth BARRICELLI(1956; see HARM1956) proposed that the of these studies was the discovery and analysis by DICK subunit theory for multiplicity reactivation be modified EPSTEINand his colleagues of the amber mutants of such that part of the phage was composed of largish phage T4. subunits (“vulnerable centers”), while the remainder The discovery of the ambers has been touched upon of the phage was composed of many small subunits. elsewhere (EDGAR1966). Thepresent history is a fuller The sensitivity of a single infecting phage would be a account, offering perspectives not previously detailed. measure of hits anywhere in it. Since one would rarely The story isvery much of the Rochester T4 Group, hit all copies of any given small subunit in any multi- headed by GUS DOERMANN,during the period from complex, the survival of the multicomplexes would be 1953 to the early 1960s. GUS’Sgroup was deeply in- determined primarily by the vulnerable centers. The volved in phage radiobiology, a discipline whose prac- quasi-final slope of the multicomplex survival curve titioners hurled poorly characterized reagents at invisi- would tell the fractionof a phage particle that was com- ble targets and hoped for interpretable responses. posed of vulnerable centers, while the shape of the LUNA (1947) and LUIUAand DULBECCO(1949) ob shoulder of the curve would provide a count of the served that bacterial cells infected by more than one number of vulnerable centers. By this analysis, 40% of UV-irradiated phage particle produced aburst of viable T4 was composed of about three vulnerable centers. progeny phage with a higher probability than expected Beginning about 1952, DOERMANNand MARTHA if the irradiated phage survived independently of each CHASE, working with T4, conducted crosses between other. This “multiplicity reactivation” demonstrated single particles of UV-irradiated phage and several par- that irradiated phageparticles could cooperate to come ticles of genetically marked,nonirradiated phage back to life. LUNA and DULBECCOoffered a simple, well (“cross-reactivation experiments”). They foundthat defined hypothesis forthe phenomenon. They pro- the fraction of multicomplexes that produced phage posed that T2 phage (a close relative of T4) are made did notfall with increasing UV doses, but that theprob- of functionally distinct subunits of equal UV sensitivity. ability that any givengenetic markerfrom the irradiated A phage particle is “killed” by UV when any one of its parent would appear in the progeny did decrease. Anal- subunits is “hit.” An infected cell will produce progeny ysis of these data showed that markers,unless very close phage if, among the several infecting particles, there is to each other, were independently eliminated from the at least one un-hit subunit of each type. yields of the individual multicomplexes (DOERMANNet DULBECCO(1952) later recognized thata “critical al. 1955). DOERMANNhoped thatthis kind of “probing” test” of the subunit hypothesis required data collected would help him in his efforts to get a fuller description at higher UV doses. At such doses, the model required of the T4 genomethan was possible using the few that thesurvival curve for multiply infected cells (multi- plaque-morphology mutantsthat wereavailable. He complexes) become exponential,with a slope the same hoped that UV lesions would serve as generic markers as that seen for singly infected cells (monocomplexes). that could be placed at high density throughout the He noted that the survival curve for multicomplexes genome. The problem, of course, was that the placing did become exponential at high dose, but the slope of these lesions was certain to be different from particle was only about 0.2 of that seen with monocomplexes. to particle, limiting their usefulness as “markers.” Accordingly, DULBECCOrejected the uniform-sensitivity SEYMOURBENZER (1955) used the rZ1 mutants of T4 subunit model. In 1956, HARMdid similar experiments to examine geneticfine structure, exploiting theinabil-

Genetics 141: 439-442 (October, 1995) 440 F. W. Stahl ity of rZZ mutants to grow on A lysogens, which do sup- spot at Caltech for a faculty job in Missouri. EPSTEIN port growth of wild-type T4. DOERMANNexploited the arrived at Caltech soon thereafter, and in late 1959 selective-growth property of these mutants to extend visited us in Eugene, where I had landed after fleeing his cross-reactivation analysis to high doses. He found from Missouri. that a wild-type allele of any conventional point muta- During the Oregon visit, EPSTEINand I discussed the tion rIZmarker was “knocked out” (ie., not transmitted state of UV radiobiology in T4 and identified a paradox. into alive phage particle) athigh doses with a sensitivity As mentioned above, rII gene function is required in A that was 1/180 of that of the plaque-forming ability of lysogens but not in nonlysogens. The lysogen in stan- a phage particle (DOERMANN1961). This high resis- dard use in GUS’Slab was the Escherichia coli K12 deriva- tance of individual markers suggested that the bit of tive K12S (A). The nonlysogen was E. coli B, the standard genome transferred from a UV-killed phage to an unir- host for T-phage experiments. From KRIEG’S work, de- radiated coinfecting phage in an individual act of scribed above, the rZI gene functions appear to be vul- “cross-reactivation” could be as small as 0.0056 of the nerable centers in K12S(A). As part of his thesis work, genome. EPSTEINhad carried out multiplicity reactivation experi- In genetically mixed infections of A lysogens, the wild ments in that strain, comparing the survival ofwild- type is dominant andphage are produced. DAVEKRIEG, type T4 multicomplexes with multicomplexes made of a student in the DOERMANN group, exploitedthis domi- acomplementing mixture of rZZA and rZZB mutants. nance of the rZZ’ allele to assess the sensitivity of gene These experiments supported theview that the rIIgenes functions to UV inactivation (KRIEG 1959). HeUV-irra- act as vulnerable centers in K12S(A). The argument diated wild-type T4 and adsorbed them to a A lysogen underlying that conclusion was laid out in a review I at low multiplicity along withseveral particles of rZZ wrote while at Caltech (STAHI.1959). However, the rZZ mutant phage. He plated the complexes before lysis on genes should not act as vulnerable centers in strain B, a host that was permissive for both rIZ and rZZ+ phage because null mutants of rIIgrow well in strain B. During in order to measure, as a function of UV dose, the his visit to Eugene, EPSTEINremarked that his recent fraction of mixedly infected cells that could produce experiments showed that the multiplicity reactivation phage. KRIEG found that the UV sensitivity for the rIZA curves for T4 in those two hosts were not distinguish- function was 10% of the sensitivity of the plaque-form- able. They should have been! The requirement for rll ing ability of T4, while for the smaller rIIB cistron the function in K12S(X) should have increased both the sensitivity was 5% of the plaque-forming ability. These high-dose slope of the multiplicity reactivation curve values were comparable to the estimated sensitivity for and the estimate of the number of vulnerable centers. a vulnerable center(40%/3 = 13%). Other experi- We realized that the paradox could be resolved by pro- ments by KRIEG had shown that, for phage production posing that T4 had two genes whose functions were in the lysogen, the rII’ function must be provided early required in B but not inK12S(A), and that thefunctions in infection. of these two genes were about as UV-sensitive as were Putting KRIEG’S radiobiological analysis of gene func- the functions of the rZI genes. tion together with DOERMANNand CHASE’Scross-reacti- Eventually, EPSTEINreturned from Oregon to Gal- vation experiments provided a semi-molecular model tech, where he shared an apartment with graduate stu- for multiplicity reactivation. This model supposed that dent CHARLEYSTEINBERG. Referring to an event in early vulnerable centers are genes specifylng early functions 1960, CHARLEY(personal communication) writes, that must be expressed before the onset of genetic re- “Dick brought up therZZmirror gene hypothesis several combination, which in T4 is so frequent that one dam- times, and I was not enamored of it. I just found it age-free chromosome can almost always be assembled difficult totake radiobiology that seriously. . . . One from damaged ones as long as the functions for doing evening at supper,with wine, he brought thehypothesis it have survived (BARRICELLI1956). This view received up yet again . . . and I said with considerable irritation, support from the thesis work of DICKEPSTEIN (1958), ‘Dick, youdon’t believe that cockamamie idea any more also a student in DOERMANN’Sgroup. DICKconducted than I do. If you did, you would have long ago started multiplicity reactivation experiments in which infection to hunt for mutants in those genes. You don’t do it was made by a mixture of two genetically marked par- because you know you won’t find any mutants.’ Dick ents. Qualitatively, an expectation of the modelwas real- was taken aback by my fury and said that he would do ized: at high dose, each productive cell gave a burst it that very night after supper. I felt morally obliged to composed primarily of one genotype, which was often help him . . .” recombinant for the markers employed. More on EP- DICKwrites (personal communication),“Charley, of STEIN’S work later. course, was an essential partner, butI do not remember BOB EDGAR,who, as a student with DOERMANN,had his encouragement to dothe experiments as being identified localized negative interference in T4, arrived angry and impatient. It isn’t Charley’s style [and] he as a postdoc in MAX DELBRUCK’SCaltech lab shortly was agreeable to picking 2000 plaques in the first try. before I (also a DOEKMANNstudent) left my postdoc . . . We managed . . . to convince Harris [BERNYI‘EIN, Perspectives 44 1

Caltech graduate student] to help us and offered the LARD BERGERand FREDEISERLING, and with LURLAand dubious reward of naming the mutants after him. Har- MARIE-LOUISEDIRKSEN at MIT. EPSTEINsoon after ris . . . hadthe nickname Immer Wieder Bernstein moved to Geneva, where he continued studies to deter- (i.e., Forever Amber) . . .” mine the stage in the life cycle at which each of his That night, several apparent &specific mutants were amber mutants was blocked. EDOUARDKELLENBERGER, isolated (“amber” mutants, of course). However, addi- who made possible the early electron microscope stud- tional mutant isolations plus complementation tests re- ies of amber-infected cells, soon exploited the muta- vealed about 20 genes rather than the two that were tions for the analysis of particle morphogenesis. At Cal- anticipated. CHARLEYwrites, “When I told Max [DEL- tech, EDGARand BILL WOOD laterconducted such BRCJCK]about all the genes we were finding, hisre- studies in vitro, with results that opened the way to the sponse was ‘How dull!’ ’’ analysis of complex assemblypathways using in vitro Obviously, the original motivation for looking for B- complementation. In Geneva, BEN HALL, PETERGEI- specific genes was no longer useful, but thereality ofan DUSCHEK, BRUCEALBERTS, and others were influential abundance of B-specificmutations was now undeniable. in initiating new biochemical studies of the mutants. DICKwrites, “. . . we fairlyquickly grasped that the For PETERand for BRUCE,contact with the amber mu- mutants might open the way to a characterization of tants led to career investigations of T4 transcription the genes of T4, and some primitive physiological stud- and DNA replication, respectively. ies , . . were among our first efforts . . .” EPSTEIN’Sand EDGAR’Sparallel studies on ambers and BOB EDGAR (personalcommunication) writes, ts mutants became more intense when BOB discovered “[When I heard of DICK’Smutants], I was filled with that his ts mutants were, for the most part, in the genes envy and wanted my own genes. So [I looked for and] that were identified by DICK’Sambers. At the 1963 Cold found the [temperature-sensitive mutants of T4]. I was Spring Harbor Symposium, the paper by DICKEPSTEIN, led to that during a conversation with [ALLAN]Camp- TOINONBOLLE, CHARLEYSTEINBERG, EDOUARD KEL bell at Cold Spring Harbor about his [host-defective LENBERGER, E. BOY DE LA TOUR,R. CHEVALLEY(Ge- mutants] and Dick’s ambm, which led us to the notion neva), and BOB EDGAR,MILLARD SUSMAN, GETTADEN- HARDT ALEX of conditional lethals , . .” EPSTEINwrites (personal and LIELAUSIS(Pasadena) introduced the communication) that it was through JEANWEIGLE that world to the awesomepower of conditional-lethal the Caltech group became aware of the possible rele- mutations (EPSTEINet al. 1964). The appearance of this vance of CAMPBELL’Swork to the understanding of the publication implied that it was plausible to undertake ambers, CAMPBELL, who was DOERMANN’Ssuccessor at a complete developmental analysis of a sophisticated Rochester, has reviewed (1993) the history of the host- biological system. defective (hd) mutants of X and of his interactions with The amber mutants and their ts cousins, found by WEIGLEand the T4 group. graduates of the DOERMANN groupRochester at as spin- A satisfactory explanation for the specificity of the off from their radiobiological analyses, provided the amber mutations was obtained by comparing the plat- phage group with generic, genome-wide markers that ing propertiesof these mutants with the plating proper- could do for phage genetics what random radiation ties of the hd (later called sus) mutants (CAMPBELL and damages could never accomplish (and what RFLPs and BALBINDER1958; CAMPBELL 1959,1961) and of “ambiv- SSRs now accomplish for human genetics). They pro- alent” dImutants (BENZERand CHAMPE 1961).Those vided a convincing demonstration of the circularity and hosts that plated hd mutants and some of those X lyso- dimension of the T4 linkage map (STAHLet al. 1964; gens on which ambivalent r11 mutants made plaques STREISINGERet al. 1964), revealed the remarkable clus- also plated the ambermutants. Apparently, KlPS(X) and tering of its genes according to function (EPSTEINet al. many other strains of E. coli could suppress the mutant 1964), andprovided the material for an elegant demon- phenotype of certain alleles of any gene (BRENNERand stration of the colinearity of a gene and its polypeptide STRETTON1964). E. coliB could not suppress the pheno- product (SARABHAIet al. 1964). More importantly, the types of those alleles. BRENNERand STRETTONdecreed steadfast pursuit of an explanation for multiplicity reac- that all such suppressible mutants be called amber. As tivation led to the discovery of mutants that freed the envisioned (dimly) by YANOFSKYand ST. LAWRENCE phage field from the genetic and radiobiological for- (1960) and(more clearly) by BENZERand CHAMPE malisms of the time by opening the door to studies of (1962), the subsequentidentification of chain-termina- development that employed direct means for analyzing tion triplets and of mutant tRNA that can read those gene function. triplets as if they stood for certain amino acids provided ALLANCAMPBELL offered helpful criticisms. DICKEPSTEIN, CHARLEY a satisfymg molecular explanation for these suppress- STEINBERGand BOB EDGARadded both accuracy and vitality through their responses to my early efforts; DICKhelped polish my final draft. ible mutants. EPSTEIN,in themeantime, had moved to UCLA, LITERATURE CITED where he undertook studies on the function ofhis vari- BARRICEIU,N. A,, 1956 A “chromosomic” recombination theory for multiplicity-reactivation in phages. Acta Biotheor. 11: 107- ous mutants in collaboration with two students, HIL 120. 442 F. W. Stahl

BENZER,S., 1955 Fine structure of agenetic region in bacterio- EPSTEIN,R. H., 1958 Ph.D. Thesis, University of Rochester, Roches- phage. Proc. Natl. Acad. Sci. USA 41: 344-354. ter, New York. BENZER,S., and S. P. CHAMPE,1961 Ambivalent rII mutants of phage EPSTEIN,R. H., A. BOLLE,C. M. STEINBERG,E. KEILENBERGEK, E. BOY T4. Proc. Natl. Acad. Sci. USA 47: 1025-1038. DE IA TOUR,R. CHEVAI.I.EY,R. S. EDGAR,M. SUSMN, G. H. BENZER,S., and S. P. CHAMPE,1962 A change fromnonsense to DENHARDT andA. I,IEIAUSIS, 1964 Physiological studies of con- sense in the genetic code. Proc. Natl. Acad. Sci. USA 48: 1114- ditional lethal mutants of bacteriophage T4D. Cold Spring Har- 1121. bor Symp. Quant. Biol. 28: 375-394. BRENNER,S., and A. 0. W. STWITON, 1964 The amber mutation. J. HARM,W., 1956 On the mechanism of multiplicity reactivation in Cell. Comp. Physiol. 64 Suppl. 1: 43-50. bacteriophage. Virology 2: 559-564. CAMPBELL,A,, 1959 Ordering of genetic sites in bacteriophage A by KRIEG,D. R., 1959 A study of gene action in ultraviolet-irradiated the use of galactose-transducing defective phages. Virology 9: bacteriophage T4. Virology 8: 80-98. 293-305. LLJRIA,S. E., 1947 Reactivation ofirradiated bacteriophage by trans- CAMPBELL,A., 1961 Sensitive mutants of bacteriophage. Virology fer of self-reproducing units. Proc. Natl. Acad. Sci. USA 33: 253- 14: 22-32. 264. CAMPBEIL,A., 1993 Thirty years ago in GENETICS:prophage inser- LURIA,S. E., and R. DULBECCO,1949 Genetic recombinations lead- tion into bacterial chromosomes. Genetics 133 433-438. ing to production of active bacteriophage from ultraviolet inacti- CAMPBELL,A., and E. BALBINDER,1958 Properties of transducing vated bacteriophage particles. Genetics 34: 93-125. phages. Carnegie Inst. Wash. Year Book, pp. 386-389. SWHAI, A. S., A. 0. W. STRETTON,S. BRENNERand A. BOILE, 1964 DOERMANN,A. H., 1961 The analysis of ultraviolet lesions in bacte- Co-linearity of the gene with the polypeptide chain. Nature 201: riophageT4 by cross reactivation. J. Cell. Comp. Physiol. 58 13-17. Suppl. 1: 79-94. STAHI.,F. W., 1959 Radiobiology of bacteriophage, pp. 353-385 in DOERMANN,A. H., M. CHASEand F. W. STAHL.,1955 Genetic recom- The Virtws, Vol. 2, edited by F. M. BURNETand W. M. STANIXY. bination and replication in bacteriophage. J. Cell. Comp. Physiol. Academic Press, New York. 45 Suppl. 2: 51-74. STAIII., F. W., R. S. EDGARand J. STEINBERG,1964 The linkage map DULBECCO,R., 1952 A critical test of the recombination theory of of bacteriophage T4. Genetics 50: 539-552. multiplicity reactivation. J. Bacteriol. 63: 199-207. STREISINGER, G.,R. S. EIX~and G. H. DENHARDT,1964 Chromo- EDGAR,R. S., 1966 Conditionallethals, pp. 166-170 in Phage and some structure in phage T4. I. Circularity of the linkage map. the M@ns of Molecular Biology, edited by J. CAIRNS,G. S. STENT Proc. Natl. Acad. Sci. USA 51: 775-779. and J. B. WATSON.Cold Spring Harbor Laboratory Press, Cold YANOFSKY,C. and P. ST. LAWRENCE, 1960 Gene action. Annu. Rev. Spring Harbor, New York. Microbiol. 14: 311-340.