Se t BAe mene TRANSDUCTION IN ESCHERICHIA COLI K-12 M. L. MORSE, ESTHER M. LEDERBERG, anp JOSHUA LEDERBERG Department of Genetics, University of Wisconsin, Madison, Wisconsin Reprinted from Genetica Vol. 41, No. 1, January, 1956 Printed in U.S.A. Reprinted from (kN ETICS Vol. 41, No. 1, January, Printed tn USA, TRANSDUCTION IN ESCHERICHIA COLI K-12! M. L. MORSE?, ESTHER M. LEDERBERG, anp JOSHUA LEDERBERG Department of Genetics, University af Wisconsin, \fadison, Wisconsin Received August 26, 1955 SYSTEM of genetic transduction has been discovered in the sexually fertile K-12 strain of Escherichia coli. This transduction is mediated by lambda, a temperate phage for which K-12 is normally lysogenic. The distinctive features of the Jambda-K-12 system include the following: (1) ‘The transductions are limited to a cluster of genes for galactose fermentation. The Gal loci are closely linked to each other and to Lp, the locus for lambda-mainte- nance. (2) The transducing competence of lambda depends on how it is prepared. Competent lambda is produced by induction of lysogenic bacteria; lambda har- vested from infected, sensitive hosts is incompetent. (3) The transduction clones are often heterogenotic, that is, heterozygous for the Gal genes which they con- tinue to segregate. Technical advantages of the lambda system include recombina- tional analysis by the sexual cycle and the availability of lysates in which nearly every lambda particle is competent. MATERIALS AND METHODS Cullures The origin and history of the Escherichia coli K-12 cultures studied have already been described (E. Leperperc 1950, 1952; LeprerperG and Leperpera 1953). The emphasis will be placed here on the Gal loci (+ = fermenting galactose; — = nonfermenting) and on the locus which controls the maintenance of lambda (/p1). The phenotypes of cultures with different alleles of 1; are as follows: Lysed by lambda Lyses Lf culture Lp culture (sensitive) yes no Lpy culture (lysogenic) no yes Lp culture immune) no no Regardless of their Lp, genotype, cultures have been found to adsorb lambda. Thus Lprt and Lp," are resistant to lysis by lambda in spite of their ability to ad- sorb the phage. In contrast with this, mutants resistant to lambda-2, a virulent mutant of lambda, are resistant because they do not adsorb either lambda or lambda-2 under the experimental! conditions used here. Media The media used include: broth, Difco penassay; agar for phage assay, Difco nutrient agar with 0.5 percent NaCl; indicator medium, EAB agar plus one percent 1Paper No. 589 of the Department of Genetics. This work has been supported at various times by the Atomic Energy Commission, Contract AT(11-1)-64, Proj. 10; a research grant (C2157) from the National Cancer Institute, Public Health Service and grants from the Research Committee, University of Wisconsin, with funds allotted by the Wisconsin Alumni Research Foundation. * Predoctoral research fellow of the National Science Foundation, 1953-34. M. I. MORSE, E. M. LEDERBERG AND J. LEDERBERG 143 sugar; minimal agar, D(O); and minimal indicator agar, EMS (J. LEDERBERG 1950). Special supplements were added where indicated. All dilutions of phage lysates were made in either penassay or nutrient saline broth, and cell suspensions were diluted in either 0.5 percent saline or penassay broth. General methods Plates and tubes were incubated at 37°C. When high cell densities were desired, broth cultures were aerated by bubbling filtered air through them. Propylene glycol monolaurate (Glyco Products Co., Inc.) at a final concentration of 0.01 percent was added to bubbled cultures to lessen foaming. Phage assays were made either in agar layer or by spreading a portion of dilute lysate with Gal~ cells on EMB galactose agar. Lysates containing lambda in high titer were prepared by two methods: (1) “Induced lambda” was liberated from lysogenic bacteria after treatment with ultraviolet (UV) (WrIGLE and Dersriicx 1951); (2) “Lytic lambda” was har- vested from sensitive bacteria infected with free lambda. The induced lambda was prepared as follows: aerated, penassay grown cultures of an Lp strain (ca. 109 cells per ml) were sedimented in the centrifuge, the broth discarded, and the cells resuspended in 0.5 percent saline. The cell suspensions (10 ml) were exposed to the radiation from a GE Sterilamp (45 seconds at 50 cm) in open petri dishes on a plat- form shaker. After irradiation the suspensions were diluted with an equivalent volume of double strength penassay broth and aerated at 37°C until maximal clear- ing was obtained. This usually required from 2 to 3 hours. To produce lytic lambda, an inoculum of induced lambda was adsorbed on to penassay grown sensitive cells. After the adsorption period the cells were sedimented to separate them from the penassay broth and resuspended in nutrient saline broth. The suspension was then aerated until maximal clearing was obtained (4-5 hours). Induced lysates have phage plaque titers of about 3 X 10° particles per ml, while lytic lysates have about 10”. Induced lambda was used in all experiments unless otherwise stated, Methods for testing for transduction In order to detect infrequent genetic changes, selective agar media were used: EMB agar for fermentation markers; EMB agar plus 100 micrograms per ml strep- tomycin for streptomycin resistance; minimal agar for nutritional markers. About 10° mutant cells in 0.1 ml broth or saline, and 0.1-0.2 ml of lysate were added to the surface of each agar plate and then spread with a bent glass rod. The plates were incubated 2-3 days before being scored. Transduction clones selected by these methods develop in a heavy background of unchanged cells. On EMB medium, negative cells grow at the expense of the pep- tone; by using sugar as well, positive clones form papillate outgrowths from the negative background. EMB agar serves as an indicator as well as a selective me- dium; isolated positive colonies are deeply colored, while negative colonies remain translucent (illustrated in fig. 3). The transduction clones were purified by the following procedure. Papillae were picked with a needle and suspended in 1 ml of sterile water. A loopful (ca. 0.001 144 TRANSDUCTION IN E. COLI ml) of this suspension was then streaked upon a portion of another plate of the EMB agar. These primary dispersals of the transduction clones were nearly always mixed. Direct picking and streaking, or spotting without any purification cannot be trusted. From the primary streaks a single colony that looked pure was picked to water and streaked as before. This operation was repeated once again, and a single colony from the last streaking was taken to represent the transduction clone. In addition to freeing the transduction clone from unchanged background cells, this method of purification may also act selectively within an unstable clone. Pick- ing apparently pure colonies leads to an overestimate of the fraction of non-segregat- ing clones. RESULTS The transductions Although a number of different loci affecting diverse portions of the genotype were tested, only genes of a cluster of loci for galactose fermentation were trans- duced by lambda lysates (Morse 1954). The Gal loci, of which about seven have been investigated thus far, are closely linked to one another (less than one percent recombination) and to Lf, the locus for lambda maintenance (LEDERBERG and LEDERBERG 1953, and unpublished). The transformation of Gal~ cells to Galt by induced lambda is illustrated in figure 1. Each papilla is a clone of galactose fermenting cells; on the area of the plate to which lysate was added, most of the Gal+ papillae are transduction clones. The Ficure 1.—The production of Galt papillae from a Gal” background of cells by a lambda lysate. Left, the control, no lysate added. Right, 0.1 ml of lysate from a Gal* culture. Some of the papillae have heen picked with a needle. M. L. MORSE, E. M. LEDERBERG AND J. LEDERBERG 145 TABLE 1 Transformation of Gal~ cultures by lysates of Gal* . Number of Gal+ papillae . . Recipient culture Lysate of: ete per al " “Control (no |g 1 mi Iyeate Gat" papillae pe . lysate)* .1 ml lysate Gal, Lp* Gal* 1.4 2 405 2.9 Gal,~ 2.4 2 2 a Galy-Lpt Gal* 1.4 20 356 2.4 Gal. 4.9 20 10 col Galy-Lpt Galt 1.4 47 394 2.5 Gals— 1.7 47 50 _ Gala Lp Gal* 1.4 4 2112 15.1 Gals 1.7 163 86 _ Gal, Lpst Galt 2.3 10 3020 13.1 : Galy- 1.7 10 18 a Gal, Lp t : Gal* 2.3 5 1296 5.6 Gal, Lp i Gal* 2.3 40 161 0.5 Gal, Lp't Gal* 1.4 29 ! 129 0.7 Gal, Lpt Galt 1.6 28 i 92 | 0.4 * The Gal™ papillae on the control are spontaneous reversals of phenotype. { Different stocks. ¢ Different experiments. quantitative relationships are illustrated in figure 2. The data can be summarized: (1) Regardless of the Lf. genotype of the recipient, transductions were obtained; (2) with each genotype the number of transductions was proportional to the amount of lysate plated; (3) Lp* recipient cultures gave 5 to 10-fold more papillae per unit of lysate than either Lp,;+ or Lp". Further, the transducing activity of lysates (which contain 10!° lambda per ml) varies according to the number of cells plated: (1) with Lpi+ Gal~ cultures there is a two-fold increase between 10%-10° cells per plate, with a plateau of maximal yield around 108 cells per plate; (2) Lp* Gal7 re- cipient cultures show about a six-fold increase over a similar range of cell platings, with the highest yield at the highest cell density. The transducing activity of lysates is specific; that is, a lysate of a Gal, culture will not transform Gal,- cultures (table 1) but Gal+ papillae were found with a Gal, culture. The specificity is extended further in that some galactose positive phenotypic reversals of a Gal~ culture give lysates with transducing activity on the original Gal, indicator (table 2). The different types of phenotypic reversals may be understood under the following hypothesis: (1) reverse mutations (Gal,- to Gal,*) yield cultures that give active lysates, and (2) suppressor mutations (Gal, Sup- pressor” to Gal, Suppressor*) will yield incompetent cultures when the suppressor lies outside the region transduced. From the data in table 1 and figure 2 the ratio of the transducing particles to the lambda particles in a lysate may be obtained. 2° recipient cultures give about one transduction per 106 lambda; Lp* recipients, one per 107. One per 10°10" lambda wil] be referred to as LFT (low frequency of transduction). 146 TRANSDUCTION IN E. COLI TABLE 2 The action of lysates of reverse mutants Numbers of Gal* papillae observed Lp* recipient culture Lysate of reversion a Control (no lysate) 0.1 ml lysate Gal,- Gal,t #1 0 648 Gale Gal. #1 10 96 Gal.” %2 6 552 Gals- Galy* #5 39 204 Galy* #8 25 291 ty ZOOOF ; . Galy LpS a] a. ° © 2000r Wd . — d i a. < lOOOr . . + Gal, Lp ° eens -L v { = Gal, P 2 3 x 10? Lambda PARTICLES Figure 2.—Proportionality between amount of lambda lysate (LFT) plated and number of papillae formed from Lp*, Lp* and Lp" Gal~ cultures. The ratio of papillae to lambda particles is 10-6 for an Lp* culture, 1077 for Lp* and Lp" cultures. Examination of the Gal* clones formed by transduction After purification the transduction clones were examined for changes at loci other than the Gal series. A number of markers were examined, including fermentative, nutritional, and phage and drug resistance mutations. The only changes at other loci were Lp* to Lpt in lambda sensitive recipients, and occasionally Lp” to Lpt in lambda immune cultures. Transduction clones from L* recipients were invari- ably Lt. To determine whether lysogeny was causally related to transduction, a recon- struction experiment was done. To a mixture of lysate and Gal- Lp* cells, Galt Lp® cells labelled with a mutant character were added to estimate the frequency of chance lysogenization in the untransformed cells in a transduction mixture. After papillae had formed, they were picked, purified, and on the basis of the differential label, divided into: (1) the inserted Galt, and (2) the transductions. The frequency M. L. MORSE, E. M. LEDERBERG AND J. LEDERBERG 147 TABLE 3 Comparison of the lysogenization of transformed and non-lransformed sensitize strains: reconstruction experiment Types recovered from mixture* of Lp* bacteria and Number of clones Percent of clones LFT lysate examined lysogenized Inserted Gal*Lac*s* 46 | 68.5 Recipient Gal~Lac~S* (non-transformed) 40 | 72.5 Transduction Gal+Lac—S* | 103 100. * 10° Gal-Lac—S", 100 Gal* Lact S*, 10° lambda particles. of lysogeny was determined in the two classes, and in the Gai~ background. Whereas unchanged Gal~ cells and the inserted Gal+ were each only 70 percent lysogenized, the transduction clones were 100 percent lysogenized (table 3). When Lp’ cultures were used as recipients, 14/112 (12 percent) of transduction clones formed were Lpt. Although the fraction is small, all previous attempts to lysogenize these cultures have been unsuccessful. The isolation of transduction clones evidently selects for these cells that have been infected with lambda par- ticles from the input lysate. The original Gal* strain and spontaneous reversions of the Gal— mutants have all been stable in ordinary culture. However, the Galt clones formed by transduc- tion are unstable for galactose fermentation as shown by the recurrence of negative and mosaic colonies (fig. 3). Despite many serial single colony isolations the galac- tose-positives continue to segregate galactose negative progeny. They behave as if Ficure 3.—EMB galactose agar plate spread with cells forma culture of a heterogenote, showing Gal*, Gal~ and sectoring colonies. 148 TRANSDUCTION IN E, COLI TABLE 4 Frequency of instability for galactose fermentation among the transduction clones Recipient cells | Unstable clones/total examined Percent unstable Gal;- Lp? 9/22 4t Gal, Lpt 40/48 83 Gal, Lp* 22/24 92 Gal; Lp 13/24 54 Lp* 20/24 ! 83 Lp 29/48 60 Gal; Lp* 6/8 75 Gals~ Lp? 28/48 58 Lpt 16/24 67 heterozygous for a single gene (or short chromosome segment) and may be desig- nated as heterogenotes. Instability among the transduction clones is quite frequent; 484 of 609 clones (70 percent) were found unstable (representative data are given in table 4). This estimate is probably low because the purification procedure acts selectively against unstable clones. The frequency of segregation has been estimated from the incidence of Gal~ in small clones of heterogenotes. The probability of segregation per bacterial division is about 2 X 107% (table 5). By repeated reisolation, however, heterogenotic lines can be maintained indefinitely. The segregants from the heterogenotic clones were examined with regard to their Gal and Lp character. Lysates of the segregants have no transducing activity on the Gal~ culture that was used as the recipient in forming the transduction clone and are therefore allelic to it. The same lysates continue to give one transduction per 108-10" lambda (LFT ratio) on non-allelic Gal~ cultures. With different recipient cultures the Lp alleles of the segregants were (1) Lpt recipient, all segregants Lpt; (2) Lp* recipients, all segregants Lpt; (3) Lp” cells, the segregants were usually Lp". In one instance, a heterogenote segregated both Lpt/Lp” and Galt /Gal~. Lysates prepared from the heterogenotes have two outstanding features: (1) instead of containing 10" lambda particles per ml, they seldom have titers higher than 5 X 10%, particularly if they originate from cultures containing few Gal~ segre- gants; (2) the number of transducing particles in these lysates is often nearly equal to the number of lambda particles in the lysate (table 6). These lysates will be re- ferred to as HFT (giving a high frequency of transduction), Transductions with lysates of heterogenotes Platings of highly diluted HFT lysate with Lp* and Lp* bacteria give a number of papillae. The number of papillae obtained with Lp* cells is, however, less than that obtained with Lpt. The lower yield with Lp* recipients may result at least in part from the loss of potential transductions through lysis of the recipient cell or of some of its early progeny. With HIT lysates it is possible to transform a large fraction of a cell population, and to observe transduction without strong selection. By adsorbing HFT lambda onto cells, diluting and plating on EMB galactose to obtain well isolated colonies M. L. MORSE, E. M. LEDERBERG AND J. LEDERBERG 149 TABLE 5 Frequency of segregation from the heterogenotes Test clones* Probability of Heterogenote segregation pert 10° au ee in | Number of Gal- cells Total cells bacterial divisionst Gal-_//Gal* 2.1t 6 | 1169 1 3 | 595 1 4 251 + 23 | 1252 3.6 | 9 1113 2 | 19 897 4.3 103 : 2750 6.6 319 1622 36.8 22 1966 2.0 0 | 237 0 Galy-_//Galt 1.5§ 11 | 323 8.1 2 176 3 8 1669 0.9 3 317 2 52 1236 8.2 0 10 0 36 | 1055 6.7 3 299 2 6 | 386 4 55 | 1965 | 3.1 * A fully grown culture in penassay broth was diluted to give about 10 cells per ml. Twenty samples of 0.1 ml were taken up in 0.2 ml serological pipettes which were supported in a horizontal position on a tray. The pipettes were incubated at 37°C for 4.5 hours. Each pipette was then blown out on to an EMB galactose agar plate, and the inside of the pipette washed with 0.1 ml of broth. The washing was added to the plate, and the inoculum spread with a glass rod. After 18 hours incubation at 37°C the number of Gal+ and Gal colonies on the plates was determined. + Using the equation 2 = 0.602r/N log N, (modified for the indicated units from Lurta and DetsrttcK 1943) where r = the number oj Gal~ segregants and NV = the clone size. The probability of segregation is also estimated by the fraction of cultures containing no segregants. 2.3 1 a= W log 5 (Po = fraction of cultures with no segregants.) 4 0 In the first experiment, using V¥ = 21° 1 4 1 2.3 a= - log 1/1/19 = 2.8 X 10% 1024 In the second experiment, using WV = 2! 3 = — flog 1/1/11 = 2.6 10°° 2 = Fong 108 11/ * t The assay plates showed this culture to have Gal*: Gal” in the ratio 106:4. Of the twenty sam- ples in this experiment, one contained only Gal*+, one contained only Gal”, and 18, both Gal? and Gal~. Only the plates that were counted are given. Nine plates were too crowded to be counted accurately. § The ratio of Gal+:Gol- in the parent culture was 128:19. The twenty cultures were distributed as follows: failed to grow, 9; contained only Galt, 1; contained both Gal* and Gal-, 10. Onc plate had approximately equal numbers of Gal*+ and Gal~ and was assumed to have come from a mixed inoculum. 150 TRANSDUCTION IN E. COLI tOOOF Lp*t Gal” ul bE . i= @ va | . 2 / 3 a 500Fr uJ hq wd 4 . a