JOURNAL OF BACTERIOLOGY, Jan., 1965Copyright © 1965 American Society for Microbiology

Vol. 89, No. 1Printed in U.S.A.

Chromosomal Aberrations Associated with Mutationsto Bacteriophage Resistance in Escherichia coli

ROY CURTISS II1

Department of Microbiology, University of Chicago, Chicago, Illinois, and BiologyDivision, Oak Ridge National Laboratory, Oak Ridge, Tennessee

Received for publication 20 July 1964

ABSTRACT

CURTISS, RoY, III (University of Chicago, Chicago, Ill., and Oak Ridge NationalLaboratory, Oak Ridge, Tenn.). Chromosomal aberrations associated with mutationsto bacteriophage resistance in Escherichia coli. J. Bacteriol. 89:28-40. 1965.-Tentypes of mutants of Escherichia coli K-12 resistant to bacteriophage T3 have been iso-lated, and several of these types have been studied genetically. Many of the /8,4,7,/8,4, 7,X, and /8,4,7,X, pro, 2 mutants were unstable, changing to complete sensitivityto T4 . The results with strains having /8,4,7,X, prol 2 mutations were compatible withthe hypothesis that this mutation caused a single break in the circular chromosomewhich prevented the normal association in the inheritance of the outside markers leu+and lac+. When sensitivity to T4 was regained, association in the inheritance of outsidemarkers was restored, and the resulting /S,7, X, prol 2 mutation behaved genetically asa deletion. The /8, 7,X, prol2 and /,4,7, X, prol 2 mutations caused positive inter-ference, inhibition of genetic recombination in regions adjacent to them, and the forma-tion of unstable partial diploid recombinants. One group of /8,4,7, X mutations did notoccur in the leu to try region of the bacterial genome. Other /8,4,7,X mutations in F-bacteria prevented the joint inheritance of the outside markers lac+ and gal+, pre-

sumably by breakage of the circular chromosome. Hfr and F+ strains with /8,4,7, x

mutations at this locus were unable to conjugate; therefore, a complete genetic analysisof the effects of this /8,4,7,X mutation could not be done.

Because Escherichia coli K-12 is sensitive to allthe T phages and can be used to study gene trans-fer by conjugation (Lederberg, 1947), it waspossible to study the genetic properties of phage-resistant mutants in this strain of E. coli. Thefollowing linkage relationships have been estab-lished for phage-resistance mutations: leu-/1,5-lac-/6 (Lederberg, 1947); strr-/Xv,mal1F (Leder-berg, 1955); gal-ura-cysB-anthranilic acid-/1, try- (Weinberg, 1960; Curtiss, unpublisheddata); and strr~-/S-/Xv,mall- (Hayes, 1957;Curtiss, 1962).

In this communication, further work is re-ported on phage-resistant mutants in E. coliK-12. A genetic analysis of /3,4,7,Xpro- and/3,4,7,X mutations has been made which indi-cates that these mutations are chromosomalaberrations. Postzygotic elimination (Nelsonand Lederberg, 1954), positive interference, andinhibition of recombination events have beenassociated with these aberrations.

1 Present address: Biology Division, Oak RidgeNational Laboratory, Oak Ridge, Tenn.

MATERIALS AND METHODS

Media. The following media were used: mediumML, containing NH4C1, 0.5%; NH4NO3, 0.1%;Na2SO4, 0.2%; K2HPO4, 0.9%; KH2PO4, 0.3%;and MgSO4.7H2O, 0.01%; and medium MA, pre-pared by adding 2X ML to an equal volume of3.0% melted agar. A carbon source at 0.5% finalconcentration and desired growth factor supple-ments were added to ML and MA. The amino acidand vitamin supplements were purchased com-merically and were added at optimal concentra-tions. Streptomycin sulfate was used at a finalconcentration of 200 ug/ml. Buffered saline con-tained: NaCl, 0.85%; KH2PO4, 0.03%; andNa2HPO4, 0.06% (when used as a diluent, gelatinwas added to a concentration of 100 Ag/ml). EMBagar contained Tryptone (Difco), 0.8%; yeastextract, 0.1%; NaCl, 0.5%; eosin Y, 0.04%; meth-ylene blue, 0.0065%; and agar, 1.3%. Just priorto pouring plates, the desired sugar was added togive a 1.0t% final concentration, and K2HPO4 wasadded to give a 0.2% final concentration. Penassayagar, Penassay broth, and L broth (Lennox,1955) were employed as complete media.

Bacteria. Table 1 lists the E. coli K-12 strainsused. Bacteria were maintained on Penassay

28

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CHROMOSOMAL ABERRATIONS IN E. COLI

TABLE 1. Escherichia coli K-12 strains

Genetic markerseStrain no. Mating Derivation

typethr lea Ti proi pro2 lac pros T| ades gal try str |met thi

x12 F- - - s + + -2 + 8 + -2 + s + _ W945x57 Hfr H + + s + + + + 8 + + + s + 3000x80 F+ + + s _ c _ c + + 8 + + + 8- + Xllxl01 F+ + + s + + + + 8 + + + S + - xl2xll4 F- s_ s + + -2 + r + -2 - r + - x12X137 F- - - r d _d -y + r + + + r + - C600xl48 F- + - r + + -2 + r - -2 - r + - x114X188 F- + - r -3 + -2 + r - -2 - r + - X148X212 F- + - r + + -g -2 r - + - r + - X148X278 F- - - r + -9 -y+ r + + + r + - C600

a The genetic markers are arranged in the order in which they occur on the chromosome. The follow-ing abbreviations are used: thr, threonine; leu, leucine; pro, proline; lac, lactose; ade, adenine; gal, ga-lactose; try, tryptophan; str, streptomycin; met, methionine; thi, thiamine; +, ability to synthesize orutilize; -, inability to synthesize or utilize; s, sensitive; and r or /, resistant. Numbers and letters forpro,-, pro2s, lac-, pro37, and galh mutations are isolation designations. The mutation Tlr confers resist-ance to both T, and T5 . All strains were nonlysogenic for X.

b W945 (Cavalli-Sforza and Jinks, 1956) was received from M. L. Morse; 3000 (Pardee, Jacob, andMonod, 1959), from N. M. Schwartz; and C600 (Appleyard, 1954), from J. J. Weigle. The omitted inter-vening steps in each derivation involved penicillin enrichment for ultraviolet-induced auxotrophicmutations, spontaneous selection of mutations to phage and streptomycin resistance, spontaneousreversions to prototrophy or ability to ferment, or a combination of these.

Strain X80 is /3,4,7,X,pro0,2 (0,4rprol,2).d Strain X137 is /3,7, X,proI,2 (4rprol.2).

agar slants at 4 C, and were transferred at 2-monthintervals.

Bacteriophages. The seven T phages (Demerecand Fano, 1945) were grown on E. coli B. Thegrowth of T3 and T7 on E. coli B was essential,since host-range mutants for /3 and /7 mutantsof E. coli K-12 made up 1 to 10% of the phagepopulation in lysates repeatedly grown on E. coliK-12. Phages X++ (Kaiser, 1955) and Xv (Lederbergand Lederberg, 1953) were grown on C600. Plkc(Lennox, 1955) and R17 (Paranchych and Graham,1962) were grown on x57 or xl01. All phage lysateswere stored over chloroform at 4 C. The tech-niques used for phage experiments were describedin the references cited above and by Adams (1959).

Isolation of phage-resistant mutants. Phage-resistant mutants were isolated on Penassay agarcontaining 0.8% NaCl by the spread-plate tech-nique described by Demerec and Fano (1945).Mutants were purified from phage by three serialsingle-colony isolations or by growing in thepresence of diluted antiphage serum, or by bothmethods.Mating procedure. Bacteria were grown to log

phase either in L broth or in appropriately supple-mented ML at 37 C with aeration. ML-growncultures were sedimented and resuspended in aminimal mating medium (pH 6.3) as describedby Fisher (1957). Matings were performed at 37 Cin a stationary 250-ml Erlenmeyer flask containinga total volume of 10 ml. F- bacteria were at a titer

of 3 X 108 to 4 X 108 per milliliter, and were alwaysat a 10-fold or greater excess in crosses with Hfrdonors.

Interruption of mating was accomplished by amodification of the procedure described by Hayes(1957). Samples from the mating mixture werediluted into ultraviolet-irradiated T6, and thenunadsorbed Ts was neutralized with T6 anti-serum. Violent agitation with a Vortex Juniormixer was sometimes used prior to T6 treatment.Recombination percentages were calculated by

dividing the recombinant titer X 100 by the titerof the donor strain in the mating mixture at thebeginning of the experiment. Recombinant titersin crosses with phage-resistant mutants werebased on the percentage of prototroph colonieswhich contained haploid recombinants exclusively(see discussion of Fig. 1). After purification ofrecombinants, unselected nutritional and fermen-tation characters were scored by replica plating(Lederberg and Lederberg, 1952). Phage resistancewas scored by streaking recombinant culturesagainst phage on EMB containing 0.1% glucose(Zinder, 1958), since even small amounts of lysiscould be detected by the red discoloration at thejuncture of the phage and bacterial streaks. Thesetests for phage resistance were always read a^ter6 to 8 hr of incubation at 37 C, because many ofthe phage-resistant mutants were mucoid andslime production made accurate scoring impossibleafter overnight incubation.

VOL. 89, 1965 29

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

TABLE 2. Mutants of Escherichia coli K-12resistant to bacteriophage T3Type of resistance No. isolated*

/................................... 65/3,X................................. 5/3,4 .......................... 25/3,4,x............. 3/3,7............................. 11/3,7,X (.or) t .......................... 9/3,4,7............................. 54/,4,7,X (¢,4r)t...................... 57/3,7,X,pro- (Orpro-)t ................ 4/3,4,7,X,pro- (k,4rpro-)t............ 11

Total ............................. 244

* The number isolated cannot be equated tomutant frequency.

t The symbol tr will be used in the text as ashorthand notation for joint resistance to T3,T7, and X. The symbol 0' indicates sensitivity toT3, T7, and X.

RESULTS

Types of phage-resistant mutants. Table 2 liststhe types of mutants obtained by selection withT3 in E. coli K-12. By contrast, in E. coli Bresistance to T3 and T4 are inseparable, so that,if resistance to X is ignored, the types /3, /3,7,and /3,7, pro- do not occur. All of the /3 and/3,X mutants had a rough colony type identicalto the parental sensitive strain. The mutants inthe other classes gave a continuum of colonymorphologies from rough to very mucoid. Ingeneral, the types showing resistance to T4 weremore mucoid than those not having resistanceto T4.

In fluctuation tests (Luria and Delbrutck, 1943)on the origin of mutants resistant to T3, largefluctuations were observed, indicating that themutants were of spontaneous origin.Those T3-resistant mutants having resistance

to T4 sometimes lost it after several transfers onslants or during selection for other mutations.The /3,4,7, /3,4,7,X, and /3,4,7,X,pro-mutants frequently changed to /3,7, /3,7, X,and /3,7, X,pro-, respectively. Most of thephage-resistant prolineless mutants were origi-nally isolated as /3,4,7, X,pro- and, althoughthey were usually stable during repeated trans-fers, they invariably changed to /3,7, X,pro-when any other forward mutation was isolated,whether it was -a-mutation to. drug5.,phage, oranalogue resistance or a mutation to auxotrophy.Very infrequently, restoration of T4 sensitivityaccompanied revision of auxotrophic mutationsto prototrophy. No proline-independent revert-

ants were detected in reversion studies with 11/3, 7,X, pro- and /, 4, 7,X, pro- mutants.

All of the phage-resistant mutants used in theexperiments reported in the following sectionswere resistant to T3, T7, and X. To simplify thepresentation of these results, the symbol q5r willbe used to indicate joint resistance to T3, T7,and X (see Table 2).

Time of entry for Hfr H genetic markers. For acomparison with data presented below, thelinkage relationships of the genetic markers em-ployed in this research are presented in Table 3.Hfr donor bacteria transfer their chromosome toF- recipient bacteria in an oriented sequentialorder. The time when a given Hfr marker is firsttransferred to F- cells is the time of entry forthat marker. The data in Table 3 are for the HayesHfr strain. The Cavalli Hfr strain, which trans-fers its chromosome in the order ade3 T6 lacleu ... , has also beea used. The distances betweenmarkers were essentially the same as found forHfr H.

E. coli K-12 has three pro loci, called pro1,pro2 , and pro3. The lac-pro3 region can becotransduced (Schwartz, 1963; Markovitz, 1964;

TABLE 3. Time of entry for Hfr Hgenetic markers*

Marker transferred Time of Distance betweenfrom Hfr H entryt markers

min

Ieu+ 8. 7 (13 )2.4

Tis 11.1 (3)4.0

pro,+ 15.1 (3)2.1

prO2+ 17.2 (2)0.6t

Ok'pro1,2 17.5 (2)0.3

lac+ 17.8 (10)0.2

pro3+ 18.0 (6)2.0

T68 20.0 (2)2.5

ade 3+ 22.5 (6)2.3

gal+ 24.8 (3)9.7

try+ 34.5 (2)

* Matings interrupted with ultraviolet-irra-diated T6 . F- strains used were X114, X137, X148,X188 X212, and x278.

t Numbers in parentheses refer to number ofdeterminations.

4 Distance between pro2+ and lac+.

30 CURTISS

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CHROMOSOMAL ABERRATIONS IN E. COLI

TABLE 4. Inheritance of +,4' marker from Hfr Ha

Control cross Experimental cross

Hfr H thr+ leu+ lac+ gal+ try+ str9 Hfr H thr+ leu+ lac+ q5,4' gal+ try- str,x114 thr- leu- lac- gal- try- strr X114 0,4r thr- leu- lac- 5k4r galh try- strr

Rela tive RelativeRecombinant class selectedb recombinant Recombinant class selectedb recombinant

frequencyc frequencyd

thr+ leu+ lac+ strr 100 thr+ leu+ lac+ ,,4r strr 100gal+ try+ strr 22.7 0, 4r gal+ try+ strr 23.1thr+ leu+ lac+ gal+ try+ strr 10.0 thr+ leu+ lac+ +,48 gal+ try+ strr 0.083

a The bacteria were grown in minimal media plus succinate and mated in appropriately supplementedminimal mating media with succinate for 60 min. The thr+ marker is closely linked to leu+ and enters0.5 min before leu+ in crosses with Hfr H.

b The markers employed for selection of recombinants are indicated in bold-face type. The recombi-nants were picked, purified, and then scored for unselected markers.

c Actual percentage of thr+ leu+ lac+ strr recombinants was 4.1%.d Actual percentage of thr+ leu+ lac+ 0,4r strr recombinants was 3.9%.

Curtiss and Charamella, unpublished data), ascan the pro2-lac region (Curtiss and Charamella,unpublished data). The pro, to lac region cannotbe cotransduced with Plkc. All pro3- mutantscrossfeed prol- and pro2- mutants (Pittard,unpublished data; Curtiss and Charamella,unpublished data), and there is no crossfeedingbetween prol- and pro2- mutants (Curtiss andCharamella, unpublished data). Hfr P4X6 trans-fers pro,+ as the first genetic marker and pro2+as the last genetic marker, whereas Hfr OR1(Curtiss, 1964a) transfers pro2+ first and pro3+last (Gurtiss and Charamella, unpublished data).Time of entry experiments with Hfr H (Table3), Hfr Cavalli, and Hfr ORI (Curtiss, 1964a),and subsequent recombinant analyses, showedthat the lac locus was between the pro2 andpro3 loci.No pro+ recombinants were obtained in crosses

between 4)rpro- or 4) ,4rpro- mutants and prol-or pro2 mutants, whereas pro+ recombinants wereobtained in crosses with pro3 mutants. All pro3_mutants cross-fed all o)rpro- and ), 4rpro-mutants. The time of entry for both the pro,+ andthe q5spro+ markers from Hfr Cavalli was about12 to 13 min. It is, therefore, concluded that thephage-resistant, prolineless mutants have adeletioni of 2 to 2.5 min of the bacterial chromo-some which includes the pro, and pro2 loci. Allprol7, pro2-, Prp.o2 , and 4,4rprol 2 muta-tions were complemented by the exogenote froma partial diploid strain in which the exogenotecomplemented the leu to pro2 region of thechromosome, but not the lac or pro3 region (Cur-tiss, 1964b). A test for cotransducibility of thepro, and pro2 loci by employing 4)rprol,2 recipi-

ents could not be done, since all o)rprol,2 mutantswere resistant to Plkc.

Genetic recombination with F- strains having,04r or 4),4rproj 2 mutations. Experiments de-signed to map the 4, r mutation with respectto other markers gave anomalous results. In E.coli K-12 the lac and gal loci are about 7 minapart (Table 3), and in crosses with Hfr H thelac+ and gal+ markers are frequently inheritedjointly. In the experimental cross between Hfr Hand a 4) r4 mutant obtained from X114 (Table 4),the relative frequency of thr+ leu+ lac+ gal+ try+strr recombinants (0.083) was less than 1% ofthe relative frequency of this same recombinantclass in the control cross between Hfr H andX114 (10.0). In contrast, the relative frequenciesof gal+ try+ strr recombinants were the same,regardless of whether the 4,4r mutation waspresent or not (Table 4). In the cross with thex114 ),4r mutant, only those recombinants whichhad inherited both the lac+ and gal+ markers fromHfr H also inherited the Hfr H ), 48 allele. Inthis cross, all of the thr+ leu+ lac+ strr and gal+try+ strr recombinants tested had inherited the0,4r marker from the F parent. Results similarto those in Table 4 were obtained in several othercrosses with the X114 4,4r mutant. All of theseresults indicated that the 0,4r mutation in X114was linked between the lac and gal loci andsomehow interfered with the joint inheritanceof the lac+ and gal+ alleles from Hfr H.Time of entry experiments with the xl 14

4)04r mutant showed that the thr+ leu+, lac+, gal+,and try+ markers from Hfr H were transferred inthe order and at the times expected for Hfr H(see Table 3). It was impossible to determine a

31VOL. 89, 1965

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

TABLE 5. Inheritance of 0k,4'pro1,2 marker from Hfr Ha

Control cross Experimental cross

Hfr H leu+ lac+ ade3+ str' Hfr H leu+ k,4spro0,2 lac+ ade3+ str8x148 leu- lac- ade3- strr X148 0,4rprol,2 leu- c,4rprol,2 lac- ade3- strr

Relative RelativeRecombinant class selected recombinant Recombinant class selectedc recombinant

frequencyb frequencyd

leu+ strr 100 leu+ 0,4rprol,2 strr 100lac+ ade3+ strr 45.1 0,4rprol,2 lac+ ade3+ strr 42.6leu+ lac+ strr 54.9 leu+ 0,48pro1,2 lac+ strr 1.67

a Procedure as for Table 4, except mating was for 90 min.b Actual percentage of leu+ strr recombinants was 11.3%.c The markers employed for selection of recombinants are indicated in bold-face type. The recombi-

nants were picked, purified, and then scored for unselected markers. Of 100 leu+ strr recombinants scored,3 were 0,4spro.,2 in genotype. Of 100 lac+ ade3+ strr recombinants scored, 4 were ck,48pro0,2 in genotype(3 of these were leu- and 1 leu+). Of 100 leu+ lac+ strr recombinants scored, 4 were +,4rprol,2 in genotype.

d Actual percentage of leu+ 40,4rpro1,2 strr recombinants was 5.4%.

time of entry for the 4,4" marker from Hfr H,since all of the thr+ leu+ and lac+ recombinantsremained resistant to T3, T4, T7, and X regard-less of the time when the mating was interrupted.Control experiments detected no lethal zygoticevents which might have accounted for theabsence of linkage between the lac and gal lociin the Hfr H X X114 0,4r cross. Reversion at thegal locus was the same in the presence or absenceof the q5,4r mutation. Reversion at the lac locuswas decreased, but still detectable, when a 4)4rmutation was present (Curtiss, 1962).Seven independently isolated F- 4),4r mutants

have been analyzed genetically in crosses withHfr H, and three have demonstrated reduction inthe joint inheritance of the Hfr H lac+ and gal+alleles. In one of these crosses with a well-markedF- strain, the 04,4r mutation caused a sharpreduction in the joint inheritance of the closelylinked ade3+ and gal+ Hfr H markers, and mostade3+ gal+ recombinants were 4),4S. This indi-cated that the 4,4r mutation in this strain wasbetween the ade3 and gal loci. In the other fourF- strains, the 4,4r mutations were unlinked tothe leu to try region. The results with 4,4r muta-tions linked between the lac and gal loci indicatedthat these mutations were associated withchromosome aberrations. This aberration couldbe a transposition, an inversion, or a single breakin the circular chromosome making it linear.In each case, one of the breaks associated withthe aberration would have to be between thelac and gal loci.A cross between Hfr H and a 4),4rpro, 2

mutant of X148 demonstrated that the ), 4rproi12mutation disrupted the normal association in theinheritance of the Hfr H leu+ and lac+ markers

(Table 5). Since only those recombinants whichhad inherited both leu+ and lac+ had also re-ceived the Hfr H 4,44pro+.2 allele, it can beconcluded that the 04 rproli2 locus is betweenthe leu and lac loci. Similar results were obtainedwith two other independently isolated F-04rpro419 2 mutants. In each case, the resultsindicated that the 4,4rproi:2 mutations wereassociated with some type of chromosomal aber-ration between the leu and lac loci.

Genetic recombination with an F- strain havinga 4)proli 2mutation. Before analyzing the data ongenetic recombination from a cross betweenHfr H and a F- 4)rprol 2 mutant, it will behelpful to discuss the results of a control cross,presented in Table 6. In this cross between Hfr Hand X278, pro2+ strr recombinants were selected,purified, and then tested for recombination oneither side of the pro2 locus by scoring for fourunselected markers.

In Table 6A, the frequency of recombinationper region is calculated. By comparing the percent recombination per region in the leu to T6segment with the per cent distance per region inthe same interval, it is evident that the amountof recombination in any region was approximatelyproportional to the length of that region, with theexception of region 4. In another control crossbetween Hfr H and X188, in which pro,+ strrrecombinants were selected, the amount of re-combination in any region was directly propor-tional to the length of that region as determinedby Hfr H time of entry experiments. In this cross,there was 16% recombination between leu and T1for a distance of 17.5%; 32% recombinationbetween T, and pro, for a distance of 38.5%;13% recombination between pro, and lac for a

32 CURTISS J. BACTERIOL.

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CHROMOSOMAL ABERRATIONS IN E. COLI

distance of 11%; 19% recombination betweenlac and T6 for a distance of 15%; and 20% re-

combination between T6 and ade3 for a distanceof 18%.

In Table 6B, the double recombinants were

analyzed by the method of Maccacaro and Hayes(1961) to determine the type of interferencepresent. Two recombination events are requiredto incorporate any Hfr H marker into recombin-ants in E. coli K-12. The data analyzed byMaccacaro and Hayes (1961) indicated that,once the first recombination event occurred, thesecond recombination event occurred more oftenin regions proximal than in regions distal to theregion in which the first recombination eventoccurred. The analysis in Table 6B indicatesthat when recombination occurred in region 4,the ratio of recombination in region 3 to that inregions 1 + 2 + 3 was 0.89. When recombinationoccurred in region 6, the ratio of recombinationin region 3 to that in regions 1 + 2 + 3 was only0.22. Thus, recombination in region 3 was asso-

ciated more often with recombination in theproximal region 4 than with recombination in thedistal region 6. Similarly, the ratio of recombina-tion in region 4 to that in regions 4 + 5 + 6 was

20 times higher when recombination occurred inregion 3 (0.60) than when recombination oc-

curred in region 1 (0.03). These results, indicat-ing that there is negative interference in E. coliK-12, are in complete accord with those ofMaccacaro and Hayes (1961).Table 7 contains an analysis of ckproT+2 strr

recombinants from a cross between Hfr H andx137. In this cross, there was 70% association inthe inheritance of the leu+ and lac+ Hfr markers.This contrasts with the almost complete absenceof leu+ lac+ recombinants in the cross betweenHfr H and the 4,4rproT,2 mutant of X148 (seeTable 5). The parental strain, from which X137was derived, was completely resistant to T4.The change to T4 sensitivity occurred concomi-tantly with the selection of a strr mutation. T4had an efficiency of plating of 0.5 on x137.The analysis in Table 7A shows that the

orproi-2 mutation in x137 has caused a significantreduction in the frequency of recombination inregions 3, 4, and 5. Based on the lengths of theregions, the amount of recombination in region 3should be 1.67 times that in region 2. Instead, theamount of recombination in regions 2 and 3 wasequal. Based on all the matings with x137, theamount of recombination in region 4 was about0.1%. This is 30 times lower than expected. Thus,on either side of the 4rproi,2 marker in x137,the frequency of recombination per region was no

longer proportional to the length of that region.It should further be noted that, in the control

TABLE 6. Analysis of pro2+ strr recombinants in acontrol cross between Hfr H and X278a

leu T1 Po02. lac T6 str

+ s + + s s

Hfr Hb _- A _

1 2 3 4 5 6

X278 ---

- r - - r r

(A) Analysis of recombination by regionsc

Recombination events in regionDetermination Total

12 3 4 5 6

Number 210 65 194 127 106 224 926Per cent 23 7 21 13.5 11.5 24 100Per centd -13 39.5 26 21.5 - 100Per cent dis- -21 54 5.5 19.5 - 100

tancee

(B) Interference analysisf

No. of recombinantswith recombination inregion 1, 2, or 3, with

Region respect to recombina- Total Ratio of 3 totion in region 4, 5, or 6 1 + 2 + 3

1 2 3

4 6 8 112 126 0.895 57 16 27 100 0.276 139 34 48 221 0.22

Total 202 58 187 447Ratio of 4 to 0.03 0.14 0.604 + 5 + 6

a Bacteria were grown and mated in L broth.Mating was interrupted after 60 min; 455 pro2+strrrecombinants were analyzed.

b Distance between markers based on time ofentry data for Hfr H (see Table 3).

c Based on double and quadruple recombinationevents.

d Per cent recombination omitting recombina-tion events in regions 1 and 6.

,,Per cent distance based on time of entryexperiments with Hfr H (Table 3).

f Based on double recombination events.

cross with X278, only 47% of all recombinationevents occurred in the outside regions 1 and 6(Table 6A), whereas in the X137 cross 71 % of therecombination events occurred in regions 1 and 6(Table 7A). Strains X278 and X137 were bothderived from the same F- strain (Table 1); there-fore, the observed differences can be ascribed to

33VOL. 89, 1965

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

TABLE 7. Analysis of 0-p?-rot strr recombinantsfrom a cross between Hfr H and an F-

with a frproi 2 markera

leut 7T1 41,pl'01,2 lac T'6 str+ s 5+ ± s s

LIuHfrHH ' I A*

1 2 3 4 5 6

X137 ___Il __V_- r r-i L- r r

(A) Analysis of recombination by regionsb

Recombination events in regionDetermination 1 -3 4 Total

1 2 3 4 5 6

Number 247 73 73 0 66 273 732Per cent 34 10 10 0 9 37 100Per centc -34.5134.5 0 31 - 100Per cent dis- - 27 45 3.5 24.5 - 100tanced

(B) Interference analysise

No. of re-combinants withrecombination inregion 1, 2, or 3, Ratio ofwith respect to

Region recombination in Totalregion 4, 5, or 6

ito 2to 3to1 2 3 1+ 2 1++2 1 + 2

+ 3 + 3 + 3

4 0 00 0 --5 49 6 3 58 0.85 0.10 0.056 171 40 43 254 0.67 0.16 0.17

Total 220 46 46 312Ratio of 4 + 0.22 0.13 0.075 to 4 +

a Bacteria grown and mated in L broth. Matingfor 60 min. Distance between markers based ontime of entry data for Hfr H (see Table 3) andHfr Cavalli; 339 k'pro0,2 strr recombinants ana-lyzed."Based on analysis of double and quadruple

recombination events. Compare with Table 6A.cPer cent recombination omitting recombina-

tion events in regions 1 and 6. Compare withTable 6A.

d Per cent distance based on time of entryexperiments with Hfr H (Table 3) and Hfr Cavalli.These percentages are different than those inTable 6A since the 4rpro1,s mutation in X137 hascaused a deletion of the pro, to pro2 segment.

e Based on analysis of double recombinationevents.

differences between the pro2- point mutation inX278 and the o)rpro-12 deletion mutation in X137.The analysis of the q5sprot,2 strr recombinants

demonstrates that the q5rprol-.2 mutation hascaused strong positive interference (Table 7B).Thus, when recombination occurred in the proxi-mal region 5, the ratio of recombination in region3 to that in regions 1 + 2 + 3 (0.05) was muchlower than this same ratio when recombinationoccurred in the distal region 6 (0.17). Likewise,the ratio of recombination in regions 4 and 5 tothat in regions 4 + 5 + 6 was least when recom-bination occurred in the proximal region 3 (0.07)and greatest when recombination occurred inthe distal region 1 (0.22). This positive inter-ference found with x137 with its orproi,2 muta-tion sharply contrasts with the negative inter-ference observed with X278, which has a pro2-point mutation (Table 6B).

All the crosses in this section were also donewith the Cavalli Hfr with essentially the sameresults. Thus, the observations cannot be ascribedto the polarity of chromosome transfer by Hfr H,since the Cavalli Hfr transfers the same chromo-somal segment in the opposite direction.To summarize this section, it can be stated that

the pleiotropic mutation, o)rproL-, causes nodisruption in the joint inheritance of outsidemarkers as does the 4,4rprO,2 mutation. The,orproi-2 mutation behaves like a deletion, sinceit does not revert and no pro+ recombinants areobtained in crosses with prol- or proi- donors.This mutation causes positive interference andalso interferes with recombination in regionsadjacent to it.

Effect of ), 4r and ), 4rpro- 2 mutations in Hfrand F+ bacteria. All 4, 4r and 4,4rpros.2 mutantswere independently isolated from various Hfrand F+ strains and had 4,4r and 4), 4rpro-,2mutations which were of independent origin fromthose in F- strains cited in previous sections.Two-thirds of the F+ 4, 4r mutants isolated failedto yield recombinants in crosses with F- bacteria(Table 8). A similar result was observed for Hfr4) 4r mutants. Two Hfr 4X,4 mutants which failedto yield recombinants were mixed with F-bacteria and observed by phase-contrast micros-copy to detect conjugating pairs (Lederberg,1956). No conjugating pairs were seen. All of therecently isolated F+ and Hfr 044r mutantswhich failed to yield recombinants have beentested for sensitivity to the donor specific ribo-nucleic acid (RNA) phage R17 (Paranchych andGraham, 1962) and found to be resistant. Pre-sumably, in this type of mutant the cell wall hasbeen altered so as to prevent conjugation. Severalof the nonconjugating F+ and Hfr 4 ,4r mutants

34 CURTISS

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CHROMOSOMAL ABERRATIONS IN E. COLI

were also mated with F- X, 4r mutants. Norecombinants were detected in these matings.Nonconjugating F+ 04) rmutants were found toact as recipients of genetic material in crosses

with phage-sensitive Hfr and F+ strains.Those F+ and Hfr 04,rmutants which yielded

normal recombination frequencies were R17-sensitive and had 4r mutations unlinked to theleu to try region of the genome (Table 8). AllF+ and Hfr 4, 4rproli2 mutants tested were

fertile (Table 8).Inheritance of the 4, 4rpro-1,2 marker from F+

donors. Table 9 presents data from two crosses

employing F+ 0),4rpro,o2 mutant donors. Incross A, the 4,4rpro0,2 F+ marker was inheritedby 59.5%70 of the thr+ leu+ recombinants. The4,4rpro-12 and lac+ markers from the F+ donorwere associated 6% of the time among thoserecombinants which had inherited the 04)4rproi,2marker (3/47). Normally, the pro2 and lac lociwere linked (see Table 6).

Cross B (Table 9) employed a 4,4rpro1,2mutant isolated from X101. Strain X101 was co-isogenic with X148 (see Table 1), and, therefore,any anomalies in linkage could be ascribed to the4,4rprol-2 mutation. Testing of the X1Ol4) rprol?2 donor culture revealed that 8% ofthe cells were sensitive to T4, and, therefore,4)rproi2 in genotype. Among the leu+ strrrecombinants the q4,4rpror2 or 4)rpro-l,2 markerwas inherited 58% of the time. Of the 45 leu+strr recombinants which had inherited the F+044rpror 2marker, only 5 were lac+, whereas of

TABLE 8. Effect on recombination frequency of¢,4r and +,k4rprol,2 mutations in Hfr and

F+ bacteria*

No. of RecombinationMating donors in frequency as

category percentage ofcontrolt

F+ X F- (control)F+ 0,4T X F- 3 50-200

3 5-2012 <1

F+ 0, 4rprol,2 X F- 4 40-300

Hfr X F- (control)Hfr ,4rX F- 4 30-180

18 < 10-4Hfr 0,4rprol,2 X F- 4 50-240

* All matings were for 60 min in broth. The 4,,4rand 0,4rproL,2 mutants were obtained from severaldifferent F+ and Hfr strains. The parent donorstrain was used as a control in each case.

t Values of <1% for F+ crosses and <10-4%for Hfr crosses indicate that no recombinantswere detected.

TABLE 9. Inheritance of the 0,4rpro1,2 markerfrom F+ donorsa

Recombinant typeb No. Per cent

Cross Acmet+ thr+ leu+ X0,4rpro7 2 lac+ 3 3.8met+ thr+ leu+ 0,,4rproj,2 lac- 44 55.7met+ thr+ leu+ c,4'pro0,2 lac- 32 40.5

Total 79 100 .0

Cross Bdleu+ 0,4rprol, 2 lac+ strr 5 5leu+ 4Orproi,2 lac+ strr 12 12leu+ q5,4rpro0,2 lac- strr 40 40leu+ cbrproi,2 lac- strr 1 1leu+ q5,4'prol,2 lac- strr 33 33leu+ +,4'prol,2 lac+ strr 9 9

Total 100 100

leu+ ck,4rpro1,2 lac+ str 1 1leu++rproi,2 lac+ strr 14 14leu- p,4rprol,2 lac+ strr 1 1leu- krpr0l,2 lac+ strr 1 1leu- c,4'pro0.,2 lac+ strr 72 72leu+ q5,41pr0l . 2 lac+ strr 11 11

Total 100 100

a The bacteria were grown and mated in broth.After 60 min for mating, the cells were centrifugedand resuspended in buffered saline before platingon selective medium.

b The markers employed for selection of re-combinants are indicated in bold-face type. Therecombinants were picked, purified, and thenscored for unselected markers.

c Cross A:x80 F+ met- thr+ leu+ 0,4rpro0,2 lac+x12 F- met+ thr- leu- 0,4'prol,2 lac-

d Cross B:X10l F+ leu+ ,,4rpro0,2 lac+ str8x148 F- leur ck,4'prot,2 lac- strr

From the X101 F+ culture used in the mating, 100isolates were tested for T4 sensitivity. Eight weresensitive to T4 and therefore q5rproT,2 in genotype.The leu+ strr recombinants were twice as frequentas lac+ strr recombinants.

the 13 which had inherited the F+ 4)rproi,2marker 12 were lac+. Among the lac+ strr recom-binants, 17% inherited either the 4,4rpro1,2 or42rp.oTi F+ marker. However, of these 17 lac+strr recombinants 15 were 4)rproj,2 and only 2 were04,rproi.2 in genotype. These results indicatedthat the F+ 0,4rproL,2 mutation interfered withthe joint inheritance of the F+ leu+ and lacemarkers, whereas the F+ 4)rpro12 mutation didnot prevent the normal association in the in-heritance of the F+ leu+ and lac+ markers. Sincethe X101 4),4rproli2 mutant was completelyresistant to T4 upon initial isolation, it is evidentthat the change in state from 4), 4rprol,2 to

VOL. 89, 1965 35

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

0 o-.14- + . ,I OL

0 20 40 60 0 20 40 60TIME (min)

FIG. 1. Time of entry for Hfr (a) and Hfr H 4k,4rprol,2 (b) markers into x2l2 (F- leu- lac- pro3- ades-strr). Bacteria were grown and mated in L broth. Mating was interrupted with ultraviolet-irradiated T6,and unadsorbed phage was neutralized with antiserum to T6 . The F- strr marker was employed in all re-combinant selections. Only 10%0 of the pro+ recombinants in the Hfr H ,,4rproi,2 X x212 cross were stable.

Orpro012 was responsible for the restoration ofthe normal association in the inheritance of theoutside markers leu+ and lac+. The data in Table9 are in complete accord with the results obtainedwith an F- 4,4rproL2 mutant (Table 5) and withan F rrpro7,2 mutant (Table 7).

Inheritance of the 4),4rpro72 mmarker from Hfr H.Two independently isolated 4),4rpro7,2 mutantswere obtained from Hfr H for use in time ofentry experiments. Both mutants were slightlysensitive to T4 (efficiency of plating of 10-i) andyielded similar results in crosses with various F-recipients. All nonrecombinant donor isolatesand recombinants which had inherited the4,4rpr.o marker demonstrated this same weaksensitivity to T4.

Figure 1 presents the data for time of entryexperiments with Hfr H (a) and with one of theHfr H 4,4rproL 2 mutants (b). The time of entryfor each marker was the same in both matingswhile the per cent recombination was twice ashigh in the Hfr H 4,4rproj,2 cross.The results with the Hfr H q5,4rrproy,2 strain

demonstrate another phenomenon observed inall crosses with 5rprol.2 I,4rproj2, and 4)4rmutants. The F- strain employed in the crosseswith Hfr H and the Hfr H 0,4rproi-2 mutanthad a prO3- mutation which was closely linkedto the q5rproL2 locus (see Table 3). Therefore,pro+ recombinants in the cross with the Hfr H

gb, 4rpro-,2 mutant should have been infrequent.There are two explanations for the high yield ofpro+ recombinants.The first explanation is that, after chromosome

transfer, the transferred partial chromosome isnot immediately integrated to form haploidrecombinants but persists as a replicating exo-genote. This replicating exogenote could then belost or integrated at a later cell division. Theper cent pro+ recombinants in Fig. lb is based onplate counts. The pro+ colonies varied in sizefrom very small to large. The analysis of pro+colonies obtained from platings after 60 min ofmating indicated that only 10% containedhaploid recombinants exclusively. All of thesewere large-colony types. Resuspending entiresmall colonies and then plating on proline-deficient media resulted in the formation ofseveral to 1 million pro+ colonies per originalcolony. The sizes of these colonies also showedgreat variation. Large-colony types, each contain-ing haploid recombinants of one genotype, weresometimes observed among the descendants fromone small-colony type. However, different haploidrecombinant genotypes were frequently obtainedfrom one original small-colony type. This resultis reminiscent of the repeated recombinationevents observed by Anderson (1958) in his studyon cell pedigrees of exconjugants in E. coli K-12.When the small-colony types were resuspended

36 CURTISS

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CHROMOSOMAL ABERRATIONS IN E. COLT

and plated on proline-deficient media, most of thecells failed to grow. By diluting the colony sus-pension and plating on Penassay agar, it wasshown by replica plating that many of the cellshad integrated the *,4rprofl, marker. Afterseveral serial small-colony isolations from proline-deficient media to eliminate any contaminatingnonrecombinant F- cells, replica plating indicatedthat nourecombinant F- cells were still present.Further serial small-colony isolations did notreduce the frequency of these nonrecombinantF- segregants. Therefore, in these instances thetransferred exogenote either did not replicateat the same rate as the F- chromosome or wasrandomly excluded from some descendants at celldivision.The formation of unstable partial diploids with

the eventual loss of the transferred exogenote byintegration or exclusion is an example of post-zygotic elimination (Nelson and Lederberg, 1954).This phenomenon was observed in all the crosseswith 6prpro- ro2, and OX4r mutants.Therefore, in all crosses with phage-resistantmutants reported in preceding sectionss, recombin-ants were purified by picking into buffered salineand then restreaked on the original selectivemedia. The procedure employed eliminated allbut haploid recombinants and stable partialdiploid strains which were sometimes obtained(Curtiss, 1964b). In this manner, it was possible toreisolate all the recombinants picked from theHfr H X X212 cross (Fig. la). In the HfrH ,4rpro0,2 X X212 mating (Fig. lb) only 72%of the leu+, 66% of the lace, and 74% of theade3+ colonies could be reisolated upon restreak-ing.The second explanation for the high number of

pro+ reoombinants in the Hfr 0,4rpr6f,2 cross(Fig. lb) accounts for almost all of the stablepros+ haploid recombinants (10% of the totalpro+ colonies). Of 99 pro,,+ haploid recombinantsanalyzed, all were 0,4$prot, and 95 were lac+.Since the distance between the 0,4rprol,2 and lacloci is only twice that between the lac and Pro3loci (Table 3), the amount of recombinationbetween the q%4'pro1,2 and lac loci was muchhigher than would have been predicted.The lac+ strr recombts from the Hfr H

*,4rpror,2 X X212 mating (Fig. lb) are analyzedin Table 10. The summary presented in Table 1OAdemonstrates that the *),4pro,°2 marker hascaused a significant reduction in recombinationin regions 3, 5, and 6, whereas there was a 10-foldexcess in recombination in region 4.The analysis in Table lOB indicates that there

was no negative or positive interference, since arecombination event in one region had no in-fluence on the randomness of the second recom-

TAi,m 10. Analysis of lac+ 8Irr recombinants froma cross between Hfr H 0,4rpro1s2 and X2126

leu T1 0,4,prol,2. lac pro3 T6 ade3 strHlfrH + s r- + + s + 8

q0,4r pro16 +L I

1 2 3 4 5&6 7 8

X212 -Tr III-

- r s+ A - r - r

(A) Analysis of recombination by regionsRecombination events in region

Determination Total1 2 3 4 5&6 7 8

Number 146 59 43 78 43 111 120 600Per cent 24 10 7 13 7 19 20 100Per cent4 - 18 13 23 13 33 - 100Per cent distance .-21l32.5 19.5 22 - 100

(B) Interference analysist

No. of recombi-nants with re-combination in

regi.on 1,2, 38 or Ratio of4,wt especttorecombintion

Region in region S and Total6,7, or 8

1 to 2 to 3 to 4 to1 2 3 4 1+ 2 1+ 2 1+ 2 1+ 2+3 +3 +_3 +3

+4 +4 +4 + 4

5&6 17 5 3 8 33 0.52 0.15 0.09 0.247 49 14 8 17 88 0.56 0.16 0.09 0.198 51 15 11 27 104 0.49 0.14 0.11 0.26

Total 117 34 22 52 225

aThe lac+ strr recombinants analyzed weretaken from platings after 60 min of mating (seeFig. 1).

b Distance between markers based on time ofentry data for Hfr H (see Table 3).

* Based on 225 double recombinants, 36 quad-ruple recombinants, and 1 hextuple recombinant.

d Per cent recombination omitting recombina-tion events in regions 1 and 8.

*Per cent distance based on time of entryexperiments with Hfr H (Table 3) and Hfr Cavalli.

f Based on double recombination events.

bination event. The lack of interference noted inthis cross contrasts with the negative interferenceobserved in the control mating between Hfr Hand x278 (Table 6) and the positive interferenceobtained in the cross with the O'proj,, mutant,x137 (Table 7). The 10-fold excess in recombina-

37VOL. 89, 1965

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

tion events between the p,4rproi,2 and lac loci(region 4) might have prevented the detection ofthe expected positive interference in this cross.

DISCUSSIONSome 4, 4r mutations in F- bacteria caused

4,ighficaIii decreases in the joint inheritance of theHfr H lac+ and gal+ markers (Table 4). It wasthus suggested that these 4,,4r mutations wereassociated with some type of chromosome aberra-tion. This aberration could be a transposition, aninversion, or a single break in the circular chromo-sunie, making it linear. The finding that donorstrains with 4,4r mutations (presumably linkedbetween the lac and gal loci) were unable toconjugate (Table 8) made it impossible to clearlydifferentiate between these alternative hy-potheses.

All F- ,)4rpro-. mutations tested caused adecrease in the joint inheritance of the Hfr Hleu+ and lac+ markers (Table 5). When the4, 4rpro-2 mutation was in F+ donors, it alsoeffectively prevented the joint inheritance of theF+ leu+ and lac+ markers (Table 9). However, thespontaneous change from 0,4rproL2 to 4,rpro-2restored linked inheritance of the leu+ and lac+markers (Tables 7 and 9). The simplest way toexplain these data with F+ and F- 4,, 4rpro- 2and 0rpro_,2 mutants is to postulate that theoriginal 04rpro 2 mutation caused a singlebreak in the circular chromosome, making itlinear, and that restoration of the circular chromo-some was accompanied by a return of T4 sen-sitivity. It is difficult to construct a model forrestoration of the original linkage if inversions ortranspositions are involved. Furthermore, oneof the two break points would have to be outsideof the leu to try region, and in one experiment withan F- ,,4rprol2 mutant the inheritance andlinkage of markers in other regions of the chromo-some was normal.

This model of a single break in the circularchromosome caused by 4,, rprot 2 mutationsreadily explains the almost complete absence ofleu+ lac+ recombinants in the crosses with F+4,4rprol 2 mutants, since both markers wouldnot be transferred to the same recipient cell.However, when the 0,4rpro,2 mutation is in theF- parent it might be expected that the trans-ferred Hfr chromosome, which contains linkedleu+ and lac+ markers, would restore the circularchromosome and give rise to normal frequenciesof leu+ lac+ recombinants. Since this is not ob-served (Table 5), it must be concluded that theF- leu and lac loci are physically separated sothat the Hfr chromosome segment cannot readilypair with both F- loci.

This model of a single break in the circularchromosome caused by 4,4rprol 2 mutationsfurther postulates that the circular chromosomeis reformed upon return of T4 sensitivity. Toexplain this (fortuitous coincidence?), it i ist beassumed that the original mutation does n -,occurin a gene responsible for some structure i volvedin T4 infection. Rather, the break in the -ircularchromosome must interfere with the functioningof a neighboring gene(s) for T4 sensitivity.Restoration of the circular chromosome couldthen allow the near-normal expression of thisgene(s).

Since the o,rproL2 mutation is a deletion(genetically), it is apparent that the originalbreak in the chromosome resulted in a loss ofgenetic material. Therefore, it is possible thatdeletion mutations in bacteria arise by a processinvolving chromosome breakage, loss of geneticmaterial and then rejoining of the free ends.The data obtained from crosses with F+

4,, 4rpro. mutants and with F- 4, 4r04 rpro -2, and orpro,,2 mutants are all com-patible with the above model. It was reasoned,however, that proof of this theory would requiretime of entry experiments with Hfr 4,4rpro-,2mutants. It was expected that the lac+, pro3+,and ade3+ markers from a Hfr H 4, 4rpro,,2donor would enter early like leu+ but be unlinkedto leu+. It is obvious from the results presentedin Fig. 1 and in Table 10 that this prediction wasnot borne out. The only justified conclusion fromthe experiments with the Hfr H 4, 4rpro012mutants was that the aberrations associated withthese 4,4rpro 2 mutations were not inversions ortranspositions.One possible explanation for the results ob-

tained with the Hfr H 4,4rproj,2 mutants wouldbe that an Hfr H strain with a linear chromosomedue to a break at the 4, 4rpro.2 locus wouldprobably be inviable, since, in Hfr H, chromosomereplication proceeds sequentially from the at-tached F (Nagata, 1963). In contrast, Nagata(1963) showed that there was no single fixedorigin for the initiation of chromosome replicationin an F- strain. Since F- and F+ 4, 4rpro0,mutants appear to have a linear chromosome witha single break at the 4,, 4rproL2 locus and areviable, it would be predicted that chromosomereplication in these mutants begins at one endof the break. Preliminary experiments supportthis, and a more rigorous test of this predictionis now being made. It is thus possible that thedifferent results obtained in crosses with F+ andF- 4, 4rpro-, mutants as compared with thoseobtained with Hfr H 4,,4rproL2 mutants aredue to a difference in the mode of chromosome

.38 CURTISS

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

CHROMOSOMAL ABERRATIONS IN E. COLI

replication in F+ and F- strains as opposed tothat in Hfr strains.

If the Hfr H 0,4rpro-2 mutants do have acircular chromosome with no break at the4, 4rrpro2 locus, then it must be explained whythese mutants are T4-resistant. As implied above,the association of T4 resistance with a linearchromosome and T4 sensitivity with a circularchromosome may be fortuitous. If the openchromosome does interfere with the functioningof a neighboring gene necessary for T4 sensitivity,it is then also possible that a larger deletion couldencompass this neighboring gene and give rise toa X5,4rprol 2 mutation which could have either alinear or a circular chromosome.

There is another possible explanation for theresults obtained with the Hfr H q5,4rproL 2mutants. Wollman and Jacob (1958) stated that,to obtain a fully fertile Hfr recombinant in across with an F- strain, it was essential to inte-grate both the proximal and the distal regions ofthe Hfr chromosome. Thus, a strain with adistally attached F but with no origin would be-have as an ineffective donor. Therefore, if, in anHfr H 4)4rproL,2 strain, the chromosome wassometimes broken at the 0,4rprol 2 locus, thenthe lac+ pro3+ ade3+ ... attached F segment,which would lack an origin, would transfer thelac+ marker to an F- recipient early but at anundetectable frequency. The chromosome frag-ment containing the origin and the leu+ marker,but with no distally attached F, also might failto be transferred. Thus, even if an Hfr H chro-mosome with a break at the 4),4rproi-2 locuscould replicate, cells containing such chromo-somes would be poor donors of their geneticmaterial.

In the cross with the Hfr H q5,4rproj 2 mutant(Fig. 1 and Table 10), the amount of recombina-tion between the 4),4rproj2 and lac+ markerswas excessive. This suggested that the Hfrchromosome could break after transfer, so thatthe lac pro3 ade3 segment could sometimes beintegrated by recombinants independently of theleu to 0,4rprol,2 region. [Recall that with F+4,4rpro7 2 donors the q5,4rpro; 2 marker wasinherited with leu+ and not lac+ (Table 9).] Ifthis explanation is correct, then it should bepointed out that it cannot be determined whetherall or only some of the transferred partial chromo-somes broke at the q6,4rpro 2 locus after transfer.Taylor and Adelberg (1961) showed, in crossesbetween Hfr donors and F- phenocopy Hfrrecipients, that markers on the proximal portionof the donor chromosome were inherited at afrequency of 70% among recombinants inheritinga marker on the distal end of the donor chromo-some. Similar results were reported by Wollman

and Jacob (1958) for Hfr X F- crosses. Thus,when an entire linear chromosome (Hfr) pairswith an entire circular chromosome (F-), thereis a linked inheritance of markers on either sideof the break in the linear chromosome. Thus, itmight be possible to obtain the observed 77%association in the inheritance of the 4),4rpro-j2and lac+ Hfr markers (Table lOB), even if all ofthe transferred partial chromosomes from theHfr H 4),4rpro-,, donor had breaks at the4),4rprol- locus. It should be pointed out thatthere is no apparent reason why chromosomebreakage should occur as a result of chromosometransfer.The above discussion indicates that the results

obtained with two Hfr H 4),4rprop9 mutants areexplanable on the basis of known mechanisms ofHfr chromosome replication and transfer. Thus,no substantial change is required in the originaltheory that 4),4rproL2 mutations cause singlebreaks in the bacterial chromosome. It is proposedthat both Hfr H 4, 4rproL 2 mutants have cir-cular chromosomes and that their q,4rpro)mutations are deletions. Thus, it would be impos-sible to obtain T4-sensitive revertants from theseHfr H 4,4rpro ,2 mutants and, in fact, no T4sensitive revertants have been found.

ACKNOWLEDGMENTS

The author expresses deep gratitude to JamesW. Moulder for his advice, support, and encourage-ment during this investigation. The generosityof M. L. Morse, N. M. Schwartz, J. J. Weigle,F. Meyer, and W. D. Fisher in supplying bacterialand phage strains is greatly appreciated.

This study was begun during the tenure of aPredoctoral Fellowship from the Division ofGeneral Medical Services, U.S. Public HealthService. The research at the University of Chicagowas supported by Graduate Training Grant 5T1GM-603 from the Division of General MedicalSciences, U.S. Public Health Service, and bygrants from the Abbott Laboratories and from theDr. Wallace C. and Clara A. Abbott MemorialFund of the University of Chicago. The researchat Oak Ridge National Laboratory was supportedby the U.S. Atomic Energy Commission undercontract with Union Carbide Corp.

LITERATURE CITED

ADAMS, M. H. 1959. Bacteriophages. IntersciencePublishers, Inc., New York.

ANDERSON, T. F. 1958. Recombination and segre-gation in Escherichia coli. Cold Spring HarborSymp. Quant. Biol. 23:47-58.

APPLEYARD, R. K. 1954. Segregation of new lyso-genic types during growth of a doubly lysogenicstrain derived from Escherichia coli K12. Ge-netics 39:440-552.

Vo L. 89, 1965 39

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from

J. BACTERIOL.

CAVALLI-SFORZA, L. L., AND J. L. JINKS. 1956.Studies on the genetic system of Escherichiacoli K-12. J. Genet. 54:87-112.

CURTISS, R. 1962. Studies on the genetics ofEscherichia coli. Ph. D. Thesis, University ofChicago, Chicago, Ill.

CURTISS, R. 1964a. An Escherichia coli K-12 strainwith the fertility factor attached to or withinthe structural gene for j8-galactosidase. Bac-teriol. Proc., p. 30.

CURTISS, R. 1964b. A stable partial diploid strainof Escherichia coli. Genetics 50:679-694.

DEMEREC, M., AND U. FANO. 1945. Bacteriophage-resistant mutants in Escherichia coli. Genetics30:119-136.

FISHER, K. W. 1957. The role of the Krebs cyclein conjugation in Escherichia coli K12. J. Gen.Microbiol. 16:120-135.

HAYES, W. 1957. The kinetics of the mating processin Escherichia coli. J. Gen. Microbiol. 16:97-119.

KAISER, A. D. 1955. A genetic study of the tem-perate coliphage X. Virology 1:424-443.

LEDERBERG, E. M. 1955. Pleiotropy for maltosefermentation and phage resistance in Escher-ichia coli K-12. Genetics 40:580-581.

LEDERBERG, E. M., AND J. LEDERBERG. 1953.Genetic studies of lysogenicity in Escherichiacoli. Genetics 38:51-64.

LEDERBERG, J. 1947. Gene recombination andlinked segregations in Escherichia coli. Genetics32:505-525.

LEDERBERG, J. 1956. Conjugal pairing in Escher-ichia coli. J. Bacteriol. 71:497-498.

LEDERBERG, J., AND E. M. LEDERBERG. 1952.Replica plating and indirect selection of bac-terial mutants. J. Bacteriol. 63:399-406.

LENNOX, E. S. 1955. Transduction of linkedgenetic characters of the host by bacteriophageP1. Virology 1:190-206.

LURIA, S. E., AND M. DELBRUCK. 1943. Mutationsof bacteria from virus sensitivity to virus resist-ance. Genetics 28:491-511.

MACCACARO, G. A., AND W. HAYES. 1961. Pairinginteraction as a basis for negative interference.Genet. Res. 2:406-413.

MARKOVITZ, A. 1964. Regulatory mechanisms forsynthesis of capsular polysaccharide in mucoidmutants of Escherichia coli K12. Proc. Nat.Acad. Sci. U.S. 51:239-246.

NAGATA, T. 1963. The molecular synchrony andsequential replication of DNA in Escherichiacoli. Proc. Nat. Acad. Sci. U.S. 49:551-559.

NELSON, T. C., AND J. LEDERBERG. 1954. Post-zygotic elimination of genetic factors in Escher-ichia coli. Proc. Nat. Acad. Sci. U.S. 40:415-419.

PARANCHYCH, W., AND A. F. GRAHAM. 1962. Isola-tion and properties of an RNA-containing bac-teriophage. J. Cell. Comp. Physiol. 60:199-208.

PARDEE, A. B., F. JACOB, AND J. MONOD. 1959.The genetic control and cytoplasmic expressionof "inducibility" in the synthesis of ,-galac-tosidase by E. coli. J. Mol. Biol. 1:165-178.

SCHWARTZ, N. M. 1963. Nature of ethyl methane-sulfonate induced reversions of lac- mutants ofEscherichia coli. Genetics 48:1357-1375.

TAYLOR, A. L., AND E. A. ADELBERG. 1961. Evi-dence for a closed linkage group in Hfr males ofEscherichia coli K-12. Biochem. Biophys. Res.Commun. 5:400-404.

WEINBERG, R. 1960. Gene transfer and eliminationin bacterial crosses with strain K-12 of Escher-ichia coli. J. Bacteriol. 79:558-563.

WOLLMAN, E. L., AND F. JACOB. 1958. Sur ledeterminisme g6n6tique des types sexuels chezEscherichia coli K12. Compt. Rend. 247:536-539.

ZINDER, N. D. 1958. Lysogenization and superin-fection immunity in Salmonella. Virology5:291-326.

AA CURTISS

on August 31, 2020 by guest

http://jb.asm.org/

Dow

nloaded from