TRANSDUCTIONAL HETEROGENOTES IN ESCHERICHIA COLI M. L. MORSE, ESTHER M. LEDERBERG, anp JOSHUA LEDERBERG Depariment of Genetics, University of Wisconsin, Madison Reprinted from Ganerics Vol. 41, No. 5, September, 1956 Printed in U.S.A. CORRECTION: Page 769, 3d paragraph, second sentence. For table 8, read table 9. For seventh line, reed eighth line. Reprinted from (EN KeTICS Vol 41, No. 5, September, 1956 Printed in USA, TRANSDUCTIONAL HETEROGENOTES IN ESCHERICHIA COLI? M. L. MORSE,? ESTHER M. LEDERBERG, ann JOSHUA LEDERBERG Department of Genetics, University of Wisconsin, Madison Received April 12, 1956 HE transduction of the Gal+ factor (ability to ferment galactose) to Gal~ mutants of Escherichia coli K12 has been described in a previous report (MorsE, LEDERBERG, and LEDERBERG 1956). The galactose positive transduction clones were often found to be unstable and to throw off Gal~ types about once per thousand divisions. We postulated that the transformed cells were heterogenotic (heterozygous for the transduced fragment), and that the instability was a result of segregation. This process has been studied in more detail with several non-allelic Gaf~ mutants. Since transduction genetics is a system analogous but not identical with sexual crossing, which also occurs in these strains, a distinctive terminology is a useful tool for integrating hypothesis and experiment. The following definitions are given for reference at this point. Their applications will be amplified in the experimental report. GLOSSARY AND SYMBOLS Genetic transduction—transfer of a genetic fragment from one cell to another. Exogenote—a chromosome fragment; usually relates to the donor in transduction. Endogenote—homologous part of the intact chromosome which corresponds to a given exogenote; usually relates to the recipient in transduction. Syngenote—(cf. synkaryon) a cell whose genetic complement includes an exo- genote (ie. is hyperploid for a fragment). Heterogenote—(cf. heterozygote) a syngenote in which the exogenote and endo- genote differ in one or more markers. Homogenote—(cf. homozygote) a syngenote in which the exogenote and endo- genote carry the same marker. Transduction clone—the entire vegetative progeny of a cell which has received an exogenote, including nonsyngenotic and syngenotic descendants. Symbols: /., Syngenotes will be given as endogenote /.. exogenote; —X Trans- duction will be symbolized as donor — X recipient. CULTURES The mutants that have been used in this study were accumulated in a variety of stocks, table 1. The Gal- mutations, as will be shown, are determined at different 1 Paper No. 619 of the Department of Genetics. This work has been supported by grants (C-2157) from the National Cancer Institute, Public Health Service, from the National Science Foundation, and from the Research Committee, Graduate School, University of Wisconsin with funds allocated by the Wisconsin Alumni Research Foundation. 2 Present address: Webb Building, University of Colorado Medical Center, Denver 20, Colorado. TRANSDUCTIONAL HETEROGENOTES IN E. COLI 759 TABLE 1 Description of principal cultures Mutant? lee. Method of Seni of Genotype! | _ Gal, W-750 | UV, EMB lactose Ft Lpt M- Lacy Vi Galz W-902 | UV, EMB galactose F- Lpt T-L-Th- Maly? Gals W-892 | spontaneous, EMB lactose | F~ Lp+ T-L-Th- Lacy Maly Gal, W-518 | UV, EMB lactose FY Lp? M~ Lacy Vi Gals W-677 | UV, EMB galactose FO LptT-L-Th Lacy Maly Xyl- Mi Ara Vi Galg W-2070 | UV, EMB galactose Ft Lp P-G- Galy W-583 UV, EMB galactose F- Lp* T-L-Tk Laci Maly Xyl- Mil- Ara Vi Galg W-1210 | UV, EMB lactose Ft Lp+ M- Lacy Vi ! F compatibility factor; M, T, L, Th, P, G: auxotrophic markers for methionine, threonine, leucine, thiamin, proline, glycine. Lec, Mal, Xyl, Mil, Ara; fermentation of lactose, maltose, D- xylose, mannitol, and L-arabinose. Vi: resistance to phages T, and T;. (LEDERBERG ef al. 1951; LEDERBERG, CAVALLI, and LEDERBERG 1952). 2 Each Gal locus in this series is referable to a unique mutational event, no recurrences having yet been identified. 3 Mal; , maltose nonfermentation, is associated with nonadsorption of lambda as a pleiotropic effect (E. LEDERBERG 1955). This effect was formerly described as Lp). but closely linked loci. These markers were also transferred to other experimental stocks by transduction. The symbols have been explained elsewhere (LEDERBERG and LEDERBERG 1953). As before Lpt, Lp" and Lp’ refer to three allelic states, re- spectively; lysogenic for the phage lambda, immune and sensitive. Where required, Lp* derivatives were obtained by irradiating Lp+ stocks with UV. Lp stocks were converted to Lp (lysogenized) by exposing them to lambda. The infected clones must be carefully purified by serial colony isolation to insure a separation of stably lysogenic from infected sensitive subclones (LEDERBERG and LEDERBERG 1953). METHODS The basic technique is the selection and classification of Gal types on an indicator medium such as EMB galactose agar. Transduction was accomplished by mixing cells of one genotype with phage from another on agar plates as described previously (Morse, ef al. 1956). Two kinds of lysates were used: HFT (high frequency of trans- duction) from syngenotic bacteria, and LFT (low frequency of transduction) from nonsyngenotic bacteria. The HFT lysates were so active that Gal* transformation was readily detected by cross-brushing them with loopfuls of Gal~ bacteria (fig. 1). Conversely, Gal~ types could be obtained by growing Galt together with the appro- priate Gal- HFT lysate, and streaking out the mixtures on EMB galactose agar. Washed suspensions of compatible, auxotrophic bacteria were crossed on minimal agar (LEDERBERG, CAVALLI, and LEDERBERG 1953). Rare galactose positive recom- binants were detected on a minimal indicator medium, EMS galactose agar (J. LEDERBERG 1950). 760 M. L. MORSE, E. M. LEDERBERG, AND J. LEDERBERG FicureE 1.—Homology or “allelism” tests. Loopfuls of HFT lambda were spotted on cross streaks of galactose negative cultures and the plate incubated at 37 C for 48 hours. The genotypes of the cultures are given at left, the source of HFT lambda at the top. Each negative culture was trans- formed to a galactose positive phenotype by HFT lambda from Gal* and non-allelic Ga/~ cultures, but was not transformed by HFT lambda from its own Gal type. Note the lower yields of papillae in the Gal, Galy interactions. The plate shown contains a modified MacConkey’s medium containing galactose instead of lactose. Routinely EMB galactose agar is used for homology tests, and with re- sults similar to those shown, but because of its better photographic qualities (better contrast) Mac- Conkey’s was chosen for illustration. EXPERIMENTAL RESULTS Characteristics of Gal- mutants The Gal~ mutants of table 1 have the common trait of a negative (translucent, near white) reaction on EMB galactose agar. The Gal+ type reacts by the deposi- tion of an opaque, near purple, stain in the colony (see fig. 3, Morse et al. 1956). Similarly, in fermentation tubes with galactose and bromcresol purple in peptone broth, the Gal~ types remain negative (alkaline, no gas) for several days, while Gait+ inocula give strong positive reactions (acid, gas) within 24 hours. Both types ferment glucose promptly. A detailed biochemical analysis of the mutants is currently under TRANSDUCTIONAL HETEROGENOTES IN E. COLI 761 way in KatcKar’s laboratory (KURAHASHI, unpublished; KALcKar ef al. 1956). Four enzymes, (1) galactokinase, (2) UDP (uridine diphospho)-transferase, (3) UDP- Gal-4-epimerase, and (4) UDP-Glu-pyrophosphatase are known to be involved in the utilization of galactose. Galy and other mutants have been found to be deficient in enzyme (2) and Galz deficient in enzyme (1). As shown in table 1, many of the Gaf~ mutants were first recognized by their modification of lactose fermentation, and detected on EMB lactose agar as stable variants of Lac~ mutabile stocks (E. LEDERBERG 1952). The mutants are indistin- guishable on EMB galactose agar but differ slightly in their effects on lactose fer- mentation; Gal; , Galy and Gal; show a weaker Lact phenotype on EMB lactose agar than do Galg and Gail; . The Gal- stocks also differ in their spontaneous revertibility to the galactose- positive phenotype, as shown by counts of papillae on confluently grown plates of EMB galactose agar. As recorded for control plates in transduction assays (tables 1, 2, 7, Morse ef al. 1956) Gai; will usually give about 40 spontaneous papillae per plate, while Gai; gives about 1 per plate of about 10" cells, and the other mutants are intermediate. The rates of mutation have not been precisely measured, and the quoted figures do not distinguish mutant clones in the inocula from plate mutations. Whether the qualitative differences in revertibility are determined by the mutant loci or by modifiers has not been explicitly tested. Furthermore, many of the ap- parent reversions may be changes at loci other than the Gal~ involved. Throughout these experiments, control platings have been stressed to minimize confusion from spontaneous reversals of phenotype. When these must be rigorously excluded, double mutants, e.g., Gal; Gal, may be used as they have not yet been found to give galac- tose-positive reversions. In addition to the distinct galactose-negative mutants just summarized, a variety of weak positive phenotypes have been noted. These are typified by “Gal;’’, which was unfortunately incorporated in an important series of multiple marker stocks of E. coli K-12, W-677 and W-1177 (LEDERBERG ef al. 1951). These lines originally carried Gal; , but when this was found to interfere with the expression of the Lac marker, a phenotypic reversal was selected. After UV treatment, the Gal; mutant was then obtained, and was chosen for further pedigrees because it did not interfere with Lac, although it later proved to be a slow positive. These stocks have evidently accumulated severa] modifiers, and crosses involving them have given such a variety of galactose phenotypes (cf. contrasting designations by WoL~tman 1953; CLowEs and Row ry 1954) that their further use, particularly where the Gal—Lp region is involved, should be discouraged. Gal; itself is not closely linked to the other Gal mutants and is not subject to transduction by lambda. Differentiation of the Gal loci by crosses and by transduction The most direct means of testing for the allelism of two Gal~ mutants is to search for galactose positive recombinations. As a necessary control on this procedure, Gal- “self crosses” (e.g., Galy by Gal; ) were also conducted wherever possible: in these crosses no Galt recombinants were detected (table 2). More extensive tests may well be expected to yield some galactose positives by spontaneous reversion, 762 M. L. MORSE, E. M. LEDERBERG, AND J. LEDERBERG TABLE 2 Self-crosses of Gal- mutants Number of recombinants Cross! Galt prototrophs® Total prototrophs® Afr M~ X FTL Tie Gal X Galy 0 4, 200 ,000 Galz X Galz 0 140,000 Galy X Gala 0 800, 000 Gale X Gals 0 160,000 Gal; X Galy 0 120,000 1 Approximately 10° cells of each parent were grown together in broth for three hours, after which time aliquots were a) plated on a minimal medium whose sole carbon source was galactose; b) di- luted and plated on a minimal medium with glucose as the carbon source. which is probably the chief limitation to the resolving power of this method. Pair- wise crosses of different Gal- gave a small fraction (about one per thousand) of Galt prototrophs, which indicates that they are different but closely linked loci (table 3). The loci Gal,, Gala, and Gal, have been most widely used, and have been test- crossed in all combinations; a few additional crosses are also listed in table 3. The loci have also been differentiated by transduction: this is shown diagrammatically in table 4. With HFT lysates, several million homology tests may be performed on a single EMB galactose agar plate (fig. 1). Therefore, HFT lysates, rather than LFT, have been used as the standard reagents of the allelism tests. The transduction data summarized in table 4 therefore permit seven loci to be distinguished, despite occasional gaps. These distinctions are entirely consistent with those of testcrosses, insofar as these have been made. About 50 additional Gal- mutants which can be transformed by various lysates have been isolated, but have not been completely analysed. Segregation from syngenotes Many of the Gal* clones formed by the transduction, Galt —-X Gal-, i.e., Galt lysate and Gal- cells, are unstable for this trait (see also Morse eé al. 1956, fig. 3 and table 4). The instability is attributed to segregation because the Gal~ clones which reappear are always the same type as the original Gal~ recipient (table 5). To generalize the test as far as possible, each Gal- segregant was picked from a hetero- genotic clone which had been isolated from an independent transductiona] event. The segregants were typed by their reactions with various HFT reagents. Double mutants are characterized by their failure to react with either of two reagents, though they respond to Gal* and other Gal- lysates. Throughout this paper, the stated for- mula has been inferred from this test, together with any other measures that may be indicated. In addition, the Gal* clones obtained by Gal; —X Galz, etc., are often unstable and can give more information on the segregation process. As before, each Gal segregant was referable to an individual transduction clone to give a composite pic- ture of heterogenote behavior. The tests of these segregants are given in table 5, TABLE 3 Inter-crosses of Gal- mutants Number of recombinants A. Cross! Gal* prototrophs® : Total prototrophs® fr M7 X FOT-L-The- Gal; X Galz li 15,000 Gal, X Gals 3 82 ,000 Galy X Galy 3 90,000 Gal, X Galg 7 102 ,000 Gal; X Galz 5 14,000 Galz X Galy 152 190,000 Galz X Galy 66 400 ,000 Galz X Gals 16 93 ,000 Gals; X Gal, 36 180,000 Galy X Galy 36 108 ,000 Gal X Galy 231 125,000 Galy X Gals 28 174,000 Galt X Galy 10 160,000 Galy X Gals 83 210,000 Gale X Galy 7 180,000 Gale X Gal 36 33,000 Galg X Gal; 248 72,000 Gale X Gal; 13 80 ,000 Galy X Gal; ! 5 110,000 Galz X Galz 12 19,000 Gal; X Galy 127 66, 000 Gal; X Gal 1 102 ,000 Gals X& Galg 130 81,000 Number of recombinants B. Crass? Gal* prototrophs Total prototrophs Galy X Galz 11 13,700 Gal, X Gals 2 1,600 Gal, X Gala 38 11,600 Gal, X Galy 2 7,600 Galy X Gals 13 7,400 Galz X Gals 3 18,400 Gal; X Galy 14 17,064 Gals X Gals 44 10,000 1 See footnote, table 2, for method. a See footnote !, table 2. b See footnote !, table 2. * These data are pooled from a number of experiments with different mating type combinations. The crosses were conducted on EMS galactose agar, where galactose-positive and negative proto- troph recombinants are scored on the same plates. These methods require further standardization before the apparent recombination fractions can be given precise quantitative meaning. 764 M. L. MORSE, E. M. LEDERBERG, AND J. LEDERBERG TABLE 4 Differentiation of the Gal~ mutants by transduction Lysates! Recipient cells Gals Gal; i 1 DQ a = gS Gait Gal, Galz Gals Galy Gale Gal; Gals | | a! is x ¢ ;++++4+ 4) | | | Gals “Ga | | | i | b++++4 0 | +tt++4H+ ++++4H1 ++tttit¢ ++4+4+ 144 t+trttt ++ 14444 + 1 A suspension of each type of recipient cell was cross-brushed against the indicated HFT lysates on EMB galactose agar. + indicates that galactose-positive transformations were observed, ~ that they were not. For Galy and Gals, only LFT lysates were available and the reactions were recorded from platings of 0.1 ml lysate plus 0.1 ml cell suspension. Three kinds of exogenote are considered: (1) carrying Gal*, (2) carrying non-allelic Gal-, (3) carrying two Gal~ loci, non-allelic with the endogenotic Gal-. The segrega- tion patterns were: Gal~ segregant type (percent) Exogenote carrying Endogenotic | Exogenatic | Recombinant types Gal* 100 . 0 0 Non-allelic Gai | 88 1t 1 Double Gal~ | 8&8 10 2 In every combination shown in table 5, the Gal~ types which are recovered corres- pond exactly to the types which entered into the transduction, or to recombinations of them. The order of frequency of the segregant types was endogenotic > exogenotic > recombinant. The Gelt clone thus behaves as if the recipient cell had received a small fragment of chromosome containing the Gal+ genes (and only these since un- related markers are not changed). Segregation from the resulting partial diploid (heterogenote) is biased, and more frequently restores the endogenotic Gal~ pheno- type. Less frequently the exogenotic Gal~ is recovered, either alone or recombined as a double Gal-. To complete the segregation pattern, stable Galt segregants should be demonstrated. These are distinguished less readily than Gal- and so have not been systematically enumerated but have been isolated incidentally to a number of experiments. Genotypic formulae for syngenotes can be written Endogenole/..Hxogenotle, €.g., Galy Gal? Galg'/cz Gal? Gal; Gals for the last item of table 5. By abbreviating the same symbols, the three kinds of syngenotes summarized above can be styled as +4 —/at+t+t, t+—-/et—-+, and ++—-/2—-—+. Until relevant data are available, no implications concerning gene sequence should be read into the formulae. The data of table 5 show no obvious difference between reciprocal transductions; TRANSDUCTIONAL HETEROGENOTES IN E. COLI 765 TABLE 5 Segregation from the syngenoles Syngenote: Number of segregants observed with: Recipient Galy Donor Gal; or Galt oo so ao Galy Galz ot Gal* Gals Galy Total endogenole/ rr exogenole endogenotic exogenolic crossover Galy Lp* Galt 9 0 0 9 Lp* Galt 33 0 0 33 Galz Lp® Galt 16 0 0 16 Lepr Gal~ 20 0 0 20 Gal; Lp* Gal> 31 0 0 31 Lpt Gal* ' 20 0 0 20 Lpr Gal* 29 0 0 29 Gals L ps Gal* 29 0 0 29 Lpt Galt 15 0 0 15 201 +(100%) 201 Galy Lp* Galz 1 0 0 1 Gals 6 1 0 7 Lpy | Galy 36 6 0 42 | Gals 18 3 0 21 Gale Lp* Gal; 14 3 2 19 Galy 9 7 0 16 Gals Lp* Gala 18 3 0 21 Gals 17 2 0 19 Lpt Galy . 16 3 0 19 Lp Galz 15 3 0 18 Gals Lp* Galy 40 1 0 41 Galy 42 1 1 44 Lpt Gal, 19 2 0 21 Galy 22 1 0 23 273 (87.5%) | 36 (11.5%) |} 3 (0.9%) 312 Gals Lp} Gal, Gal 135 14 3 152 Gal; Gal; 29 1 0 30 Galy Lp* Gals Galy 18 2 1 21 Gal; Gala 12 4 0 16 194 (88.5%) | 21 (9.6%) 4 (1.8%) 219 1In these transductions, involving three factors, Galz refers to both mutant loci taken together; the crossovers are other combinations. nor is there any effect of the Zp allele of the recipient on the segregation pattern of the resulting syngenotes. As noted previously (Morse ef al. 1956) all transduction clones derived from Lt recipients were lysogenic; from Lp* recipients, Lpt+ or Lp’; and from Lp’ recipients, usually Lp’. The segregational behavior of Lp alleles is still under study and will be taken up elsewhere. The tests for Gai type were based on the pattern of transduction by lysates from known Gal~ cuJtures. As additiona] checks on the transductional test for Gal type 766 M. L. MORSE, E. M. LEDERBERG, AND J. LEDERBERG TABLE 6 Examination of segregants from heterogenotic clones | Classification! of segregant by: Heterogenote Transduction Lysate Testcross to: Gal type Gal type Endogenotic Gai- Exogenotic Gal~ Gal* total Galt total Galz /e2 Galz 1. Galz Galy 0 4070 2.' Galz Galz 0 5384 3. Galz Galz 0 2072 4, Galy Galy 0 6988 Gal fic Galt 1. Galy Gala 0 896 2. Galy Gal, 0 918 3. Gal, Gal, 0 1134 4. Galy Galy 0 863 Galz Galt lex GalfGaly 1. Galy Galy 0 2786 3 3183 2, Gal, Galg 0 2675 2 3471 3. Gal, Gal, 0 3485 23 5342 4. Gal, Galy 0 5952 1 1665 5. Galy Galy 0 5000 1 891 6. Galy Galg 7 3102 0 1988 7. Galy Gala 10 4364 0 1187 Gal; Gal? /.2 GalfGalz 1. Gal, Gay 0 16104 0 1389 2, Gals Gal, 0 5730 1 164 3. Galy Gal, 0 3358 0 202 4, Gals Gal, 0 12848 1 171 5. Galy Galz 1 11200 0 827 6. Galg Galy 6 10608 0 718 7. Galy Galg 3 5000 0 409 1 Classifications: (1) by transduction, exposure to lysates of known Gal-; (2) lysate activity, lysate of the segregant on known Gal-; (3) testcrossing with known Gal~ cultures, the figures given are Gal*+ prototrophic recombinants and total prototrophic recombinants, scored as per table 3B. a number of segregants were tested further by two methods: (1) by making lysates of them and plating the lysate with known Gal- cultures, (2) test crossing against known Gal- cultures. These further checks on the classification gave perfect agree- ment with each other and with the transductional test. Cultures of 79 segregants were checked by transduction to known Gal-, and 26 segregants were testcrossed. The scope of the analysis is indicated by the data in table 6, in which data on segre- gants tested by both methods are recorded. The sequence of events in segregation from a single heterogenote could, in principle, be studied in cell pedigrees, but the rate of segregation, about 10-* per division, would make this too laborious an enterprise at present. A number of individual heterogenotes have, however, been studied intensively by plating methods. For example, W2869, Gal, Galt / eGalp Gal was replated after purification, and no more than one Gal— segregant tested per segregating Gal+ colony. This insures the uniqueness of each TRANSDUCTIONAL HETEROGENOTES IN E. COLI 767 segregant. From a total of 112 colonies, the segregant types were 85 Gal; ; 26 Gal and 1 Gal; Gal;, which are endogenotic, exogenotic and recombinant, respectively. Syngenotes are distinguished from haploids (nonsyngenotes) primarily by their genetic complexity, which is revealed by segregation. A second distinction, with technical import for further analysis, is their behavior as transductional donors. As noted previously (Morse et al. 1956) the lysates from heterogenotic cultures show a very high frequency of transduction. Every Lpt heterogenote which has been tested has given a lysate showing HFT behavior, i.e., not less than one transduction per hundred plaque-forming particles, and often approaching one per one. Clones which give HFT lysates will be referred to as HFT*. As will be noted later, HFT+ behavior has served in turn as an auxiliary criterion for homogenotes, in which segregation is less readily observed. HFT lysates from a number of double heterogenotes, e.g., Galy Gal /xGal} Galz , have been assayed on various Gal~ types, with the results depicted in table 7. The exogenote is preferentially included in the phage which matures in the syngenotes. In view of the incidence of non-syngenotic segregants, which are known to generate much larger bursts of LFT phage than syngenotes, the true efficiency of phage from heterogenotes is systematically underestimated. The estimation of preferential in- clusion of exogenotes is also complicated by the incidence of syngenotic crossovers, as described in the following section. Homogenotic segregants Three a priori possibilities for the constitution of Gal~ segregants may be con- sidered (the first two of which have been realized): (1) reduced haploids; (2) unre- duced syngenotes, which have become homogenotic for one or more Gal- loci; (3) diploids homozygous for one or more Gal- loci (these might arise by secondary non- disjunction). If segregants occur whose genotypes consist only of exogenotes, we assume they would be inviable. Each of the events might or might not be preceded by crossing over. Two methods are available for distinguishing homogenotic (Gal-/..Gal-) from haploid, nonsyngenotic (Gal-/) segregants. Reverse mutation from Gal- to Galt would convert a homogenote to a segregating heterogenote, Gal—/..Gal- to Gal+/ exGal~ or Gal~/,,Gal*. A haploid Gal~ would revert toa haploid Gal*+, which would not segregate for galactose fermentation; this supposition is supported without exception by tests of at least 100 reversions from type Gal- cultures, usually obtained as papillae from control platings in transduction assays (see, e.g., tables 1, 2, 7, MorsE et al. 1956). In addition, by analogy with heterogenotes, homogenotic cultures should be HFT*. Haploid segregants like all of the type haploid stocks which have been tested should be LFTt. Altogether, 77 segregants from 18 different heterogenotes were screened initially by the reversion test. For each test from one to ten independent reversions was selected from each Gal~ segregant clone. Two of the 77 (2.6%) gave segregating reversion clones and are tentatively considered to be homogenotic. Both segregants happen to come from W2869 (Gal; Galz / Gal} Galz ) and are themselves Galz / eGals homogenotes. From one of these clones, both of two reversions were unstable; from 768 M. L. MORSE, E. M. LEDERBERG, AND J. LEDERBERG TABLE 7 Lysates of heterogenotes Heterogenote Assay of lysate of heterogenote for: Ratio exo: endo Endogenote Exogenote Phage? Exton Endoepte Galy Galg 7200 | 120 1 120 Gal, Galy 1 i 1.8 0.06 30 Gals; Gal, 620 150 | 43 3.5 Galz Galy 7300 25 ! 0.28 | 89 Galg Galy 21 | 14 1.2 11 1 Plaques on Lp* culture. 2 Papillae on Lpt endogenotic Gal~ culture. 3 Papillae on Lpt exogenotic Gal culture. the other, one reversion was unstable, the other stable. The last observation may be attributed to a suppressor mutation in another region, further segregation in the homogenotic clone, or to the lesser frequency of Gal~ (exogenotic) segregants from the potential Gal+/.. 44 o s[ 3 (donor) lytic (recipient : , Inet fective | crowth p ) Heterogenote Gal,” Goly” Lp* Lambda Gal, Galt tpt x" yt, ox xty7 PEt ts FicureE 2.—Diagrammatic representation of the lambda-EZ. coli transduction system. The fate of the prophage (Lp) segment has not been worked out: this is indicated by the broken outline in the diagram of the heterogenote and by the omission of this segment from figures 3 and 4. transduced are the Gal markers. Since it is precisely these markers which are linked to Lp (in crosses, LEDERBERG and LEDERBERG 1953; in transduction, Lennox 1955 and Jacos 1955), we conclude that the effective exogenotes have a special relation- ship to lambda prophage. The simplest view of this relationship is diagrammed in figure 2, that the Gal markers and the prophage (Z*) are adjacent segments of the bacterial chromosome, and that an effective particle is one in which induction has accidentally released a fragment which includes the Gal region, as well as the unique Lp segment. From this point of view, lysogenization, or the transfer of the prophage is itself a form of transduction, in which the behavior of Lp (prophage) is analogous to that of Gal. The unique features of the Lp segment which relate it to virus are the capacity for autonomous growth of Lft in the lytic cycle, and in “inducibility”, ie., the formation of specific coats in the maturation of phage when Lt bacteria are exposed to UV. The envelopment of the Gal-Lp segment in the coat of a mature phage particle protects this material during its extracellular existence, and provides the means for its reentry into a new bacterium (assuming an analogy of lambda with 7:, HERSHEY and Cuase 1952). How these ends are accomplished with free DNA is known only on a complex, empirical basis. At this point of our narrative, the exogenote has penetrated the recipient bac- terlum. This bacterium may be lysogenic, immune, or sensitive. Undoubtedly a fraction of transduction clones that would otherwise be detected from sensitive re- cipients are lost by lysis; how it is determined whether an infected bacterium will be lysed, lysogenized, or otherwise protected, is unknown. The criterion for effectiveness in transduction is the formation of a galactose- positive clone, and therefore for the functional and reproductive integrity of the Gal segment. Ineffective particles may have exogenotes that were defectively prepared, or being intact, are improperly established in the recipient cell. The high activity of TRANSDUCTIONAL HETEROGENOTES IN E. COLI 775 HFT lysates shows that establishment is not a general limiting factor; the LFT lysates may be assumed to contain a heterogeneous population of particles, of which only a few contain an effective exogenote. The HFT quality of the lysates from syngenotes may be explained on the basis of selection of effective exogenotes, which had been screened by a prior selection for the ability to form a persistent galactose- positive heterogenote. The induction of a syngenote evidently preserves the integrity of the exogenote, as shown by its preferential inclusion in the phage yield, and perhaps by the reduced yield of phage from such a cell. Two types of transformed clones have already been noted, stable and segregating. The stable clones evidently result from an early exchange of the exogenote with the recipient, leading to a reduced haploid product. The mechanism of this exchange cannot be readily studied (LEDERBERG 1955), but may be compared to the crossing over and reduction observed with persistent heterogenotes. However, the incidence of primary stable transformed clones (about }4 of the total) is much higher than would be expected from the known rate of segregation in heterogenotic clones (10-° per bacterial division, Morse e¢¢ al. 1956). It has also been observed that stable trans- duction clones are much less frequent in HFT transductions. The discrepancy might be explained either by postulating that the initial heterogenote is inherently unstable, or that LFT lysates include a moiety of particles that are capable of effective ex- change, but not of initiating a persistent heterogenote. No data are yet available on the distribution of fragment sizes; the exogenotes so far studied (in HFT lysates from syngenotes) appear to encompass all the Gal markers. More information about the mechanism of exchange can be obtained from the study of persistent heterogenotes. The three basic modes of division are diagrammed in figure 3. Most often, the heterogenote is propagated as such (mitosis); it may be reduced to a haploid, with or without concomitant crossing over; it may undergo internal recombination or automixis, i.e., engender a new syngenote of different con- stitution, for which one or more crossovers are implied. The types of automictic syngenotes already imply that crossing over occurs at a stage at which either the endogenote or the exogenote (presumably both) is already duplicated, i.e., at a “four-strand” stage. Proof that crossing over in a specific instance involves all four strands would entail, for example, the identification of a + +/.2 + ++ homogenote from a -+ —/ez — + heterogenote. Suspected occurrences of such types have not been completely analyzed (table 10). Without more explicit information on centromere relationships, it is difficult to decide whether automixis is meiotic (reductive separation of centromeres) or mitotic (equational separation). Since a diploid homozygote has not been found as a product of automixis, the latter is preferred (see figure 3 for some of the postulated expecta- tions). The term ‘‘somatic segregation” has been employed for analogous situations, but the issue here is not whether segregation occurs in a germinal or a somatic cell (whatever this would mean in the present context) but the underlying chromosome mechanics (STERN 1936). The relationship of reduction to automixis is obscure. They may be concurrent, or as in Aspergillus nidulans (PONTECORVO, GLOOR, and ForBES 1954) the two processes may be independent, reduced crossover haploids being derived from syngenotes that 776 M. L. MORSE, E. M. LEDERBERG, AND J. LEDERBERG i and i 7a FA R. Ore Qt Mitosis PL gdh) Te NY e ’ a 3 Reduction Ficure 3.—Diagrammatic representation of modes of cell division occurring in syngenotic clones M, mitosis; A, automixis; R, reduction. had undergone prior automixis. On this notion, reduction would result directly from the loss, nonreplication or nondisjunction of the exogenote. The prevalence of the endogenotic class of segregant is consistent with this view, but with most other for- mulations as well. This question may require a detailed pedigree analysis of a segre- gating clone before a definite answer is possible. Attempts to induce segregation of heterogenotes, e.g., with UV (cf. effects on heterozygotes, LEDERBERG ef al. 1951) have been unsuccessful. In figure 3, the exogenote is depicted as synapsed with the homologous endogenote. Specific synapsis must occur, at least to account for automictic events, but this is only one of several possibilities. On the one hand, the exogenote might be firmly at- tached to the endogenote, e.g., at the Zp site, to form a short “branch”; on the other it might lie free in the cell during vegetative growth. The stability of heterogenotes in vegetative growth can be explained by assuming either 1) a definite association, synapsis or attachment, of the two elements through- out the division cycle or 2) regular equational disjunction of an independent exogenote which would behave, in effect, as a separate chromosome. Automictic events would necessitate at least an occasional synapsis even on the second view. Whether it is transient or persistent, synapsis is a crucial mid-step of the trans- duction process. Transduction might be more prevalent than is now recognized, in the sense that exogenotes are transferred, but unless they retain the capacity for specific synapsis, they may be unable to attain a functional relationship with the cell. For example, the failure of lambda generally to mediate transduction of loci TRANSDUCTIONAL HETEROGENOTES IN E, COLI ~T ~~ “IT s e Popp oH e Oo e oO ® . Gis e 2. _.% —+ _—_— § ° e e @o Cis 1 ON AL UN ARPeEAPrPS AP SAP e 3. _f —+ @#o 0 0 @ Trans © 4. 4 —+f — t @o Oo Oo © 0®@ Trans Donor Recipient Heterogenote Ficure 4.—The Gal; , Galy interactions (position effect). The structural formulae, inferred from the origins of the clones and from subsequent segregational behavior, are shown diagrammatically. The phenotypes are: @, galactose positive; dotted ©, galactose negative, papillating; O, galactose negative. remote from Lf might be explained in these terms, rather than by their failure to be enveloped in the lambda particle. To put the burden of specific synapsis on the Lp segment is a plausible extrapolation from the regularity with which lambda com- bines with an L/* cell to form a lysogenic Lpt. However, other phages such as P1 do mediate generalized transduction in #. coli K-12. Pi and lambda might differ either in the quality of the nuclear fragments that are produced during viral growth (in regard to any critical step of transduction) or to the ability of the maturing phage coats to discriminate between prophage particles and other residues of the bacterial nucleus. The uniqueness of each of the seven Gal~ markers that has been studied here paral- lels previous studies of the complexity of genetic loci. The uncovering of this com- plexity depends on the adequacy of recombination tests. In maize or Drosophila, 100,000 or a million tests would be considered exhaustive. The procedures described in this paper (HFT transduction and HFR crossing) could be extended to 10° or 10 with comparable effort (compare crossing in maize, 105, LaucHnan 1955; LFT 778 M. L. MORSE, E. M. LEDERBERG, AND J. LEDERBERG transduction, 10°, DEMEREC ef al. 1955; recombination in phage, 10°-108, BENzER 1955). The numerical comparisons should not be taken too seriously, as the most serious limitation is not the efficiency of the screening methods, but the interference from other factors, especially spontaneous mutation. Microorganisms have recently been the most prolific source of “pseudoallelism” but there is no reason to suppose that other organisms differ more profoundly than in limitations of technique. Previous studies on bacteria have shown the prevalence of pseudoallelism and po- sition effect for other loci in E. coli K-12. For example, heterozygotes of the trans type Lacj,Lac},/ Lact Lach, proved to be lactose-negative, while both single hetero- zygotes were lactose-positive (E. LEDERBERG 1952). Unfortunately, the techniques then available did not allow the positive identification of the cis heterozygote, Laci, Lach,/Laci,Laci, which was assumed to be lactose-positive. In the Gai heter- ogenotes, the cis and érans arrangements have been compared directly (fig. 4), but they are yet to be studied in heterozygous diploids. SUMMARY The transduction of genetic material between cells of E. coli by phage lambda has been studied further. The material that could be transduced is limited to markers for the fermentation of galactose. Most of the transductional clones resulting carry a chromosomal] fragment and are diploid for the genes which had been transduced. Such partially diploid clones, which have been named heferogenotes, segregate about once per thousand cell divisions, and give HFT*+ lysates after UV induction. With the exception of the gene combinations, Gal; Gal} /Gal} Gal; heterogenotes of the form +/—, + —/— +, — —/+ 4, etc., were galactose positive, indicating that the + alleles are dominant. In the exceptional cases, érens combinations (+ —/-— +) had a negative, the cis (— —/+ +) a positive phenotype, indicating a position effect between these loci. The heterogenotes undergo three modes of division: (1) mitosis, to propagate the original heterogenote; (2) reduction, to form haploid clones with markers of either the original recipient cell, transduced fragment, or recombinants of the two; (3) automixis, to form a new heterogenotic clone. This mode is analogous to somatic or mitotic crossing over. LITERATURE CITED ALEXANDER, H. E., and Grace Lerpy, 1951 Determination of inherited traits of H. influenzae by desoxyribonucleic acid fractions isolated from type specific cells. J. Exptl. Med. 93: 345-359, ALEXANDER, H. E., and W. Repman, 1953 Transformation of type specificity of meningococci. Change in heritable type induced by type-specific extracts containing desoxyribonucleic acid. J. Exptl. Med. 97: 797-806. Baron, L. S., SaMuEL B. Format, and WaLTeR Sprtman, 1955 Vi phage-host interactions in Salmonella typhosa. J. Bacteriol. 69: 177-183. BEwzER, SEyMouR, 1955 Fine structure of a genetic region in bacteriophage. Proc. Natl. Acad. Sci. U.S. 41: 344-354, Crowes, R. C., and D. RowLEy, 1954 Some observations on linkage effects in genetic recombina- tion in Escherichia coli K-12. J. Gen. Microbiol. 11: 250-260. DEMEREC, M., I. BLomsTRaNnp, and Z. E, DemerEc, 1955 Evidence of complex loci in Salmonella. Proc. Natl. Acad. Sci. U. S. 41: 359-364. Epwarps, P. R., Betty R. Davis, and W. B. Cuerry, 1955 Transfer of antigens by phage lysates with particular reference to the 1,w antigens of Salmonella. J. Bacteriol. 70: 279-284. TRANSDUCTIONAL HETEROGENOTES IN E. COLI 779 EPHRUSSI-TAYLOR, HARRIET, 1955 Current status of bacterial transformations. Advances in Virus Research 3: 275-307. Grirritn, F., 1928 The significance of pneumococcal types. J. Hyg. 27: 113-159. Hersuey, A. D., and Marra Cuase, 1952 Independent functions of viral protein and nucleic acid in growth of baeteriophage. J. Gen. Physiol. 36: 39-56. Horcuxiss, Roun D., 1955 Bacterial transformation. J. Cellular Comp. Physiol. 45 Supplement 2: 1-22. Isext, S., and T. SaKar, 1954 Transduction of biochemical properties in Salmonella E group. Proc. Japan Acad. 30: 143-147. Jacos, Francois, 1955 Transduction of lysogeny in Escherichia coli. Virology 1: 207-220. Katcxar, HERMAN M., EizaBeTu P, ANpERsoN, and Kurt J. IssELBACHER, 1956 Galactosemia, a congenital defect in a nucleotide transferase: A preliminary report. Proc. Natl. Acad. Sci. U.S, 42: 49-51. LAUGHNAN, JouN R., 1955 Structural and functional bases for the action of the A alleles in maize. Am. Naturalist 89: 91-104. LEDERBERG, EsTHER M., 1952 Allelic relationships and reverse mutation in Escherichia coli. Genetics 37: 469-483. 1955 Pleiotropy for maltose fermentation and phage resistance in Escherichia coli K-12. Genetics 40: 581-582 (Abstract). LEDERBERG, ESTHER M., and Josnua LEDERBERG, 1953 Genetic studies of lysogenicity in Esch- erichia coli, Genetics 38: 51-64. LEDERBERG, J., 1950 Isolation and characterization of biochemical mutants of bacteria. Methods in Medical Research 3: 5-22. 1952 Cell genetics and hereditary symbiosis. Physiol. Rev. 32: 403-430. 1955 Recombination mechanisms in bacteria. J. Cellular Comp. Physiol. 45 Supplement 2: 75-107. LEDERBERG, J., E. M. LEDERBERG, N. D. ZINDER, and E. R. Livery, 1951 Recombination analy- sis of bacterial heredity. Cold Spring Harbor Symposia Quant. Biol. 16: 159-174. LEDERBERG, Josuua, L. L. Cavarir, and Estuer M. LEDERBERG, 1953 Sex compatibility in Escherichia coli. Genetics 37: 720-730. Lennox, E. S., 1955 Transduction of linked genetic characters of the host by bacteriophage P1. Virology 1: 190-206. Morse, M. L., EstHer M. LEDERBERG, and JosHua LEDERBERG, 1956 Transduction in Escherichia coli K-12. Genetics 41: 142-156. PonTECcORVO, G., ELIzABETH Tarr GLoor, and E. Forges, 1954 Analysis of mitotic recombina- tion in Aspergillus nidulans. J. Genetics 52: 226-237. STERN, CurRT, 1936 Somatic crossing over and segregation in Drosophila melanogaster. Genetics 21: 625-730. WEICLE, J. J.. and M. Desricx, 1951 Mutual exclusion between an infecting phage and a car- ried phage. J. Bacteriol. 62: 301-318. Wo.tmay, E. L., 1953 Sur Je determinisme génétique de la lysogénie. Ann. Inst. Pasteur 84: 281- 293. ZinDeER, N. D., and Josnua LepERBERG, 1952 Genetic exchange in Salmonella. J. Bacteriol. 64: 679-699.