Dec. Vol. In Parental Functions During Conjugation Escherichia … · macromolecular syntheses, or both, in one or both parents. These experiments were done to determine which parent

BACTERIOLOGICAL REvIEws, Dec. 1968, p. 320-348Copyright © 1968 American Society for Microbiology

Vol. 32, No. 4, Pt. 1Printed In U.S.A.

Parental Functions During Conjugation inEscherichia coli K-12

ROY CURTISS III, LEIGH J. CHARAMELLA,1 DONALD R. STALLIONS, AND JANE A. MAYS2Biology Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830

INTRoDucTioN................................................................ 320EXPERIMENTAL APPROACH...................................................... 321MATERIALS AND METHODS..................................................... 321Media................................................................... 321Bacterial Strains............................................................ 322Mating Procedures.......................................................... 322Controls ................................................................... 322Reproducibility of Experiments................................................ 324

RESULTS........................... 324Effects of Inhibiting Energy Metabolism in One or Both Parents Before and During

Conjugation .............................................................. 324Effects of growth conditions and starvation.................................... 324Effects of starvation for an energy source prior to and during mating.............. 327Function of energy metabolism in each parent during conjugation................. 327

Effects of Inhibiting DNA Synthesis Before and During Conjugation................ 330Recombinant production in matings with thymine-requiring parents.............. 330DNA synthesis and initiation of chromosome transfer........................... 330

Effects of Inhibiting Protein Synthesis Before and During Conjugation ...... ......... 331Effects of amino acid starvation on recombinant production...................... 331Effects of completion of vegetative chromosome replication on recombinant produc-

tion ................................................................... 331Effects of Genetic Inhomology Between Donor and Recipient Chromosomes on Chro-

mosome Transfer........................................................ 336Effects of deletions in the F- chromosome on recombinant production by Hfr donors.. 337Effects of deletions in the Hfr chromosome on recombinant production............. 340Effects of deletions in the F- chromosome on F' transfer........................ 341

SUMMARY DISCUSSION......................................................... 344LITERATURE CITED............................................................ 346

INTRODUCTION

A partial understanding of the mechanismsoperative during conjugation between Hfrdonors and F- recipients of Escherichia coliK-12 has been achieved as a consequence ofstudies performed in numerous laboratories.Jacob and WoUlman (39) and Hayes (35) haveevaluated and summarized many of the earlierstudies. Recent efforts to further elucidate thenature of events during conjugation have beenprimarily concerned with testing the mechanismsfor chromosome transfer proposed in 1963 byBouck and Adelberg (8) and by Jacob, Brenner,and Cuzin (38). Bouck and Adelberg proposedthat Hfr cells had to complete their present cycleof vegetative chromosome replication prior toinitiating transfer of their chromosomes without

'Present address: Department of Microbiology,University of Cincinnati, Cincinnati, Ohio.

2Present address: School of Medicine, VanderbiltUniversity, Nashville, Tenn.

further replication. Jacob, Brenner, and Cuzinpostulated that replication of the donor chromo-some during transfer was obligatory and acted asthe driving force for transfer. They also pro-posed that this chromosome replication duringtransfer was initiated following a contact stimu-lus received from the F- parent, was controlled bygenes present in the integrated fertility factor F,and was independent of any system for control-ling chromosome replication during vegetativegrowth. Adelberg and Pittard (2) have includedin their recent review a detailed discussion ofthese two proposed mechanisms for chromo-some transfer. Most investigators who haveattempted to differentiate between these twomodels have obtained results which are at vari-ance with the predictions of the Bouck-Adelbergscheme but in accord with the expectations ofthe Jacob-Brenner-Cuzin hypothesis (4, 6, 24,25, 33, 34, 36, 49).

Jacob, Brenner, and Cuzin (38) reiterated, as

320

on Septem

ber 1, 2020 by guesthttp://m

mbr.asm

.org/D

ownloaded from

http://mmbr.asm.org/

CONJUGATION IN E. COLl

part of their model, Fisher's (28) conclusionthat the F- parent is passive during chromosometransfer. Bonhoeffer (7), Freifelder (30), and thepresent authors (20; Curtiss, Mays, and Stallions,Bacteriol. Proc., p. 55, 1967) have obtained datawhich suggest that the F- parent may perform avery active role during chromosome transfer.This manuscript contains evidence we have usedin postulating a mechanism for chromosomemobilization and transfer which requires theactive participation of both mating partners.

EXPERIMENTAL APPROACHIn the experiments reported in this paper, we

used donor and recipient mutants unable toutilize various energy sources or to synthesizevarious metabolites. We conducted matings inthe presence or absence of these substances as ameans of controlling energy metabolism ormacromolecular syntheses, or both, in one orboth parents. These experiments were done todetermine which parent must be capable of per-forming which biosynthetic activities during theinitiation and continuance of chromosome trans-fer. Before discussing the experiments conductedand the interpretation of the results obtained, weshould mention certain features of our experi-mental approach to understanding the functionsof each parent during bacterial conjugation.

First, we subdivided bacterial conjugation intofive steps and attempted to study them experi-mentally one at a time under conditions whichare optimal for all steps other than the onebeing studied. The five steps are (i) formation ofspecific pairs between donor and recipient cells,(ii) conversion of specific pairs to effective pairsor conjugation tube formation, (iii) chromosomemobilization or the initial events in the donorcell for the conversion of a circular chromosomeinto a chromosome capable of being sequentiallytransferred, (iv) chromosome transfer, and (v)integration of the transferred donor chromosomeinto the recipient chromosome to producerecombinants. Brinton and his collaborators (9-11) have shown that F pili on donor cells areabsolutely necessary for specific pair formation.We (R. Curtiss and L. G. Caro, Bacteriol. Proc.,p. 27, 1966; manuscript in preparation) have con-firmed this conclusion and have discoveredgrowth conditions for donor parents whichmaximize the mean number and length of F piliper cell. We have also shown that specific pairformation occurs at unaltered frequencies in theabsence of energy metabolism in either parent(22). Therefore, all of the experiments conductedduring the past several years which are reportedin this manuscript were done under conditionsin which essentially all donor cells form specific

pairs with one or more recipient cells. Since theexact nature of conjugation tube formation isprobably not understood (Curtiss and Caro, inpreparation), it is possible that the metabolicactivities required during the inception of matingare needed for conjugation tube formation orchromosome mobilization, or for both.The second feature of our experimental ap-

proach to understanding bacterial conjugation isthe use of recombinant formation in the recipientas an indication of chromosome transfer in mostexperiments. It is preferable in some instances touse indicators of chromosome transfer which donot require integration of the transferred materialinto the recipient chromosome, and this was donewhen possible. However, such indicators oftransfer as enzyme synthesis, episome replication,and zygotic induction of a prophage all requireenergy metabolism and macromolecular synthesesin the recipient for detection. The initial recom-bination event(s) both in phage (56, 57) and inbacteria (19) probably can occur in the absenceof energy metabolism. Therefore, we believe thatrecombinant production is the best genetic indi-cator of chromosome transfer in studies in whichenergy metabolism or macromolecular synthesesare being inhibited in the recipient parent.The third feature of our experimental approach

is the use of isogenic donor and recipient strains,derived in our laboratory from the prototrophicF+ strain W1485 (43). We were prompted to dothis by the realization that most donor andrecipient sublines of E. coli K-12 have not sharedthe same ancestor for over 20 years. Thus, byestimating the number of cell division cyclesneeded for the introduction of each individualmutation and for routine transfers during these20 years, we estimate that the existing represen-tatives of these sublines are separated by about10,000 cell division cycles of growth. This isroughly equivalent to about one-quarter millionyears in human evolutionary time.

MATERIALS AND METHODSMedia

The formulas for minimal liquid and minimalagar media have been described (18). The mini-mal mating medium contained the same totalconcentration of salts as minimal liquid medium,except that K2HPO4 and KH2PO4 were added at4.9 and 6.3 g/liter, respectively, to achieve afinal pH of 6.3. Glucose was used as the carbonsource unless otherwise indicated. All carbonsources were used at a final concentration of 5.0g/liter. Difco Casamino Acids were used as asupplement to minimal medium in some experi-ments. There was no detectable difference be-tween use of solutions of Casamino Acids which

VOL. 32, 1968 321

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL.

had been treated with activated charcoal andthose which had not. Supplements were pur-chased from Calbiochem (Los Angeles, Calif.)and were used at the following concentrations(in,ug/ml): L-alanine, 100; L-aspartic acid, 100;L-asparagine, 100; L-arginine HCI, 22; L-cysteineHCI, 22; glycine, 100; L-glutamic acid, 100; L-glu-tamine, 100; L-histidine HCI, 22; L-isoleucine,20; L-leucine, 20; L-lysine HCl, 88; L-methionine,10; L-phenylalanine, 20; L-proline, 30; DL-serine,100; DL-threonine, 80; L-tryptophan, 20; L-tyro-sine, 20; L-valine, 20; adenine, 40; uracil, 40;thymidine, 4; thymine, 4; thiamine HCI, 2; andstreptomycin sulfate, 200.

Difco Penassay broth and agar, L broth andL agar (44), and EMB agar (19) were used ascomplex media. Buffered saline with gelatin (18)was used in some experiments.

Bacterial StrainsThe bacterial strains are listed in Table 1 with

their derivation. All mutations either were ofspontaneous occurrence or were induced by lowdoses of ultraviolet irradiation or by low con-centrations of nitrous acid. The enrichment pro-cedures for isolation of auxotrophic mutants andHfr donors have been described (5, 21). Allstrains were frozen soon after isolation for long-term storage. Working stocks were maintainedon Penassay agar slants at 4 C.

All strains used to determine the requirementsfor energy metabolism and macromolecular syn-theses during chromosome mobilization andtransfer possess mutations which are not leakyand which either do not revert or revert at lowfrequency. Strains X 584 and X 724 have thy-mutations which do not revert and are not tem-perature sensitive. These strains, isolated byenrichment with aminopterin, will grow slowlywith 1 ,ug of thymine/ml and at optimal rateswith 4 jug of thymine/ml. When these strains areresuspended in thymine-deficient medium afterbeing washed once, there is a 5 to 10% increasein deoxyribonucleic acid (DNA) which occursduring the first 20 min after removal of thymine,as measured by the diphenylamine reaction.Thereafter, there is no detectable net synthesis ofDNA.

Mating ProceduresDuring the course of these studies, the proce-

dures for conducting matings have been markedlyimproved. The following format has been usedduring the past 2 or 3 years, and exceptions tothese procedures are listed in the table footnotesand figure legends. Bacteria are grown for 8 to10 generations in tubes (25 X 200 mm) contain-ing L broth or containing minimal medium sup-plemented with 0.5% (w/v) Casamino Acids.

Donor strains are grown without aeration toachieve maximal numbers and lengths of F piliper cell (Curtiss and Caro, in preparation), andrecipient strains are grown with aeration at 37 C.Sedimentation, washing, resuspension, and star-vation are all conducted at 37 C to minimize anytemperature shocks. Great care is taken withdonor cultures to avoid breakage of F pili. Thematings are conducted in 10-mI volumes in 125-ml microfernbach flasks (Bellco Glass, Inc.,Vineland, N.J.) which are immersed up to within5 to 10 mm of the metal caps in a water bathmaintained at 37 C. The donor-to-recipient cellratio is usually between 1 :10 and 1:20. The totalbacterial density is no more than 2 x 108/ml atthe commencement of mating to avoid problemsof oxygen depletion toward the end of matingsof long duration. Interruption of mating isaccomplished by diluting samples of the matingmixture into ultraviolet (UV)-irradiated (2.5 x103 ergs/mm2), purified T6 bacteriophage at 2 X1010 particles/ml, final concentration. This sus-pension is then immediately agitated on a vortexmixer for 15 sec to separate mating partners.Following 15 to 16 min at 37 C, antiserum to T6,prepared by injecting rabbits with purified T6, isused to neutralize any unadsorbed T6. Afteranother 15 to 16 min at 37 C, appropriate dilu-tions are plated on streptomycin-containingmedia selective for the desired class of recombi-nants.The integration of transferred donor genetic

material into the recipient chromosome to yieldrecombinants requires a complex series of events.Thus, it is essential to provide conditions whichwill allow the F- parent to perform all necessarymetabolic functions associated with integrationto obtain maximal recombinant yields. Therefore,the minimal medium used for the T6 treatmentand all subsequent dilutions was appropriatelysupplemented to provide glucose at a final con-centration of 0.5% (w/v) and all other metabo-lites required by the F- parent. For matings in Lbroth, the minimal medium used as diluent con-tained 10% L broth (v/v).

ControlsSince Hfr strains differ with respect to the

stability of F integration (1, 12), we routinelydetermined the frequency of Hfr versus "non-Hfr" donors in the cultures used for matings byusing the cross streak methods outlined by Bergand Curtiss (5). Reisolation of pure Hfr cultureswas sometimes necessary. Hfr ORI (Curtiss,Bacteriol. Proc., p. 30, 1964) is very unstable,and cultures descended from single colonies con-tain between 60 and 90% Hfr cells. We showed,however, by analyzing the mating type of recom-binants inheriting a marker transferred early by

322 BACrRIOL. REV.

on Septem


mbr.asm

.org/D

ownloaded from


CONJUGATION IN E. COLI

Hfr OR1, that the "Fe type" donors in the Hfr mined by cross streaking log-phase recombinantOR1 population were not contributing more cultures across a donor-specific phage on EMBthan a few per cent of the recombinants formed. agar containing 0.1% glucose (5).] The recombi-[The mating type of recombinants was deter- nant frequencies in matings with Hfr ORI were

TABLE 1. Bacterial strains"

Strain Mating type Relevant genotypeb Derivation

prototroph X- strrthi- X-T6' str'iht X+ T6' str' malA+ malB+thi- X+ T6 str' malA- malBTthr leu- thi- X7 T6' strr malA+ malB+thr leur thi- X- T6 str7 malA- malE-thr leu- proA-BT thi- X- T6r str7prototroph lacZ X- T6' str'mer x T6' str'thr- leu- proC- thi X- T6& str7thr- leu- proBT thi- X7 T6r str7prototroph X- T6' str'mer thr X+ T6' str'lac- thi- x- T6& str,,IF-lac+proB-lac- X T6' str'prototroph X- T6' str'his- X- T68 str'prototroph X7 T6' str'prototroph xylt X- T6' str'purE pyr' his- xyt- X T6r str'proB-lac- X7 T6' str'proB-lac- thyr X T6& str'lacYl T6')X7 str'/F-lac+ proB+ proA+lacYl T6' - str'/F-lac+thr+ leu- proA- thi X- T6r str'proBT leu- arg- X7 T6r strrleu- arg- X- T6' strrlac- proB- leui arg- X- T6' str'proA- leu- arg- X- T6r str7proA-BT leui arg- X- T6' strrthr purE- pyr- his- xyt- X- T6r str'thr- purFr pyr his- xyt- thyr X- T6r str'lacZ-I- leu- arg X T6' str'purE- pyr- trp his- xylf malA- X- T6r str'proB-lac- leu- arg- X7 T6' strrpurE- pyr trpj his- xyt' malkA X- T6' sir'

prototroph xyt- X- T6' str'thr- proC- purE- pyr- his- xyl)X7 T6r str'thr- proA- purE- pyr- his- xyl X- T6& str'thr- leu- pro- mer thi/ T6& str'/F-thr+ leu+

x 15c300"

X 57

X 91

x 12dx 12"C600c d

x 15c. ex llcC600c, d

C600c dx l5c. fAB23839AB785 = 200PShx 289x 42cx 42cx 15c, s

x 15c i

x 289x 354x 354x 535'x 629;C600cW d

x 289x 696x 696x 705x 705x 540x 723x 705 X Plkc (1485-1k)x 540

x 705 X Plkc (x 354)x 775 X Plkc (x 289)x 503x 723x 723CB069'

a The abbreviations and nomenclature used follow the proposal of Demerec et al. (27) with the ex-

ceptions noted by Curtiss (19).b All mutations conferring auxotrophic requirements are listed, but some mutations conferring re-

sistance to drugs and phageand inability to utilize carbon sources have been omitted for sake of brevity.c Curtiss (17).d Curtiss (18).e Curtiss (19).f Curtiss et al. (21).g From K. Szende by way of N. Schwartz.k From F. Jacob by way of E. Adelberg and N. Schwartz.Berg and Curtiss (5).Curtiss and Stallions (22).Schwartz (52).

From B. Low by way of C. M. Berg.

x 18x 57x 91x 96x 99x 100x 137x 225x 235x 277x 278x 289x 313x 314x 354x 435x 436x 493x 503x 540x 545x 584x 646x 647x 680x 696x 705x 708x 710x 711x 723x 724x 733x 775x 784x 800x 801x 820x 821x 927

F+Hfr HHfr HHfr HF-F'F-Hfr OR1Hfr CavF-F-F-Hfr P4X6F'F-Hfr OR6Hfr OR7Hfr OR11Hfr OR21F-F-Hfr OR41F'ORF-206F'ORF-207F-F-F-F-F-F-F-F-F-F-F-F-Hfr OR21F-F-F' KLF-1

323VOL. 32, 1968

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL.

therefore calculated by using the actual titer ofHfr cells.

In all experiments in which one or both parentswere starved for a carbon source or metabolite,the mutation causing the inability to utilize orsynthesize the compound was checked for rever-sion by plating undiluted samples of the parentculture on appropriate selective medium. Theparent cultures were also checked for inability togrow on the media used for recombinant selec-tions.

In crosses between nonisogenic parents, thecolonies which arise on selective medium are notalways due to the presence of haploid recombi-nants (18). Therefore, we frequently picked andanalyzed colonies formed on selective media todetermine haploidy and purity of recombinanttype. This was usually done the first time a par-ticular mating was performed. Data from matingsin which problems of partial diploidy, impurityof recombinant clones, or distorted linkage wereobserved are not included in this manuscript.These problems are of infrequent occurrence inmatings between isogenic parents.

Reproducibility ofExperimentsThe experiments cited in this communication

were initiated in August 1961 and completed inApril 1968. Each type of experiment was per-formed by two or more of the authors and witha great variety of bacterial strains. Most of thedata obtained from matings with nonisogenicparents have been omitted. The experimentalresults obtained with different substrains of E.coli K-12 and by different authors were essen-tially similar for the same type of experiment.

RESULTSEffects ofInhibiting Energy Metabolism in One or

Both Parents Before and During ConjugationEffects of growth conditions and starvation.

Fisher (28) proved that zygote formation was anendergonic process requiring the aerobic utiliza-tion of an energy-yielding carbohydrate. He con-cluded from results obtained in experiments withstarved or unstarved parents that only the Hfrdonor required an energy source during chromo-some transfer. The first three lines of Table 2

TABLE 2. Effects of aeration during growth and starvation on recombinant production byHfr and F- parents grown in brotha

Growth conditions of 3-hr starvation of

Hfr parent F| parent Hfr parent F- parent

(1) Aerated

(2) Aerated

(3) Aerated

(4) Static

(5) Aerated

Aerated

Aerated

Aerated

Aerated

Static

No

Yes

No

No

No

No

No

Yes

No

No

Glucosepresentduringmating

YesNo

YesNo

YesNo

YesNo

YesNo

Frequency of thrt led' str'recombinants

Actualb Relative

10061

2.3 (100)0.73 (32)

133 (100)58 (44)

118 (100)77 (65)

18 (100)12 (64)

8.3 X 10-25.1 X 10-2

1.9 X 1036.1 X 10-4

1.1 X 10-14.8 X 10- 2

9.8 X 10-26.4 X 1-2

1.5 X 10-29.6 X 10-3

a x 57 (Hfr H thr+ leu+ strs) and x 99 (F- thr leu- strr) were grown for 3 hr in 20 ml of Penassay brothin 200 X 25 mm tubes either with (aerated) or without (static) aeration to a cell density of about 2 X108/ml. The bacteria were then sedimented, washed, and resuspended, either in buffered saline, if theywere to be starved, or in minimal mating medium containing threonine, leucine, and thiamine, if theywere to be mated. The 20-ml cultures starved in buffered saline with aeration were likewise sedimentedand resuspended in minimal mating medium before mating. Glucose was added as indicated to themating flask containing the F- parent 30 sec before adding the Hfr parent. Matings lasted for 40 minand were interrupted by dilution into minimal medium containing thiamine, glucose, and streptomycin,immediately followed by agitation on a vortex mixer for 60 sec. Appropriate dilutions were then platedon selective medium.

b Averages for three or more matings in which the inhibitory or stimulatory effects on recombinantyield were always in the same direction but differed in magnitude.

324 BAcTERIOL. REV.

on Septem


mbr.asm

.org/D

ownloaded from


CONJUGATION N E. COLI

contain mean data from experiments similar tothose performed by Fisher. It is evident thatstarvation of broth-grown Hfr cultures in buff-ered saline results in a marked decrease inrecombinant yield, regardless of whether an

energy source is present or absent during mating(line 2, Table 2). This nonrestorable loss ofmating ability is primarily due to the fact thatdonor cells lose F pili at an exponential rateduring starvation in buffered saline (Curtiss andCaro, Bacteriol. Proc., p. 27, 1966; manuscriptin preparation) and consequently lose the abilityto form specific pairs with F- cells.

Starvation of broth-grown F- cells in bufferedsaline does not decrease the actual recombinantyield in matings conducted in the absence of an

energy source (line 3 versus line 1, Table 2) aswas noted by Fisher (28) However, starved F-

cultures consistently yield higher recombinantfrequencies than do unstarved F- cultures inmatings conducted with an energy source present(line 3 versus line 1, Table 2). [This behaviorwas also observed for amino acid-requiring F-

strains starved for a required amino acid (seebelow) and is in accord with the observation thatstationary phase F- cells give higher recombinantfrequencies than do log-phase F- cells.] Thus, tocorrect for the increased recombinant yieldfound in matings with starved F- cells conductedin the presence of glucose, it is necessary to

determine relative recombinant yields. Whenthis is done, it is found that starvation of the F-parent prior to mating results in fewer recombi-nants (44%) than when the F- is not starved(61%). These effects of starvation on the F-parent are more pronounced when a distallytransferred Hfr marker is used for recombinantselection (data not shown, but see experimentspresented later in manuscript).The data in line 4 (Table 2) demonstrate that

Hfr donors grown without aeration yield more

recombinants than do donors grown with aera-

tion. The relative increase in recombinant yieldfor Hfr donors grown without aeration is morepronounced when the mating is of shorter dura-tion than 40 min or when selection is for a dis-tally transferred Hfr marker (data not shown).This behavior is probably accounted for by theobservation that donor strains grown withoutaeration have a higher mean number of F phiper cell than do donor strains grown with aera-

tion (Curtiss and Caro, in preparation). Thiswould, therefore, allow a higher frequency ofspecific pair formation. The data in line 5 (Table2) indicate that the ability of the F- parent toyield recombinants can be affected by the condi-tions of growth. Hfr and F- strains grown with-out aeration are less and more affected, respec-

tively, by starvation in buffered saline than are

TABLE 3. Effects of starvation for an energy source prior to and during mating on recombinantproduction and zygotic induction by Hfr and F- parents grown in minimal mediums

Time (min) of starvationfor glucose Glucose Actual frequency ofb Relative frequency of

present

HIfrparent F parent mating thr kteu sir' X infective | thr leu+ strr X infectiverecombinants centers recombinants centers

0 0 Yes 1.4 X 102 1.2 X 10-2 100 100No 7.8 X 103 6.9 X 10- 56 58

0 80 Yes 1.3 X 10-' 1.1 X 102 93 (100) 92 (100)No 2.1 X 10- 1.2 X 103 15 (16) 10 (11)

80 0 Yes 3.6 X 103 3.8 X 10-3 26 (100) 32 (100)No 1.7 X 103 1.8 X 10-3 12 (47) 15 (47)

80 80 Yes 2.6 X 10-' 2.1 X 10-3 19 (100) 18 (100)No 2.6 X 104 1.5 X 104 1.9 (10) 1.3 (7.1)

a X 91 (Hfr H thr+ leu+ X+ str") and x 100 (F- thr leu- X7 strr Xr) were grown with aeration in ap-propriately supplemented minimal medium containing glucose to a cell density of about 2 X 108/ml.The bacteria were then sedimented, washed, and resuspended either in supplemented minimal mediumlacking glucose, if they were to be starved for glucose, or in minimal mating medium containing thre-onine, leucine, and thiamine, if they were to be mated. The matings were terminated after 40 min. Otherprocedures are described in the footnote a of Table 2. x 18, a streptomycin-resistant derivative of theF+ strain x 15, was used as the X indicator strain.

b Average of two matings.

325VOL. 32, 1968

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL. BACTERIOL. REV.

t-

I~tCD 0<, td e0n (-Cd. - -

;;c ;;m ; m

r-r-1- r- C14 00~ 'rOOW0

C4 e

4

00

xx-

,1

eXX

-4

x xN C1

- tf)

ft m

8 8

xxO Cl

.C--

0s0Nc 0- N 0N-CD Q-,0-O - 8e

-1-

W )t ti en mtnm1_)

I.0 ~~

Xx

cl -

O C)_

XX" O.4tr t

Xxel ll

ra m

8 8_4 I"

Xx00 -

C- '-

m m

XXx -C1 CS

C4m

8 8xxO

-4 4

C0C-0 0 0 0 0 0 0 0 0

QCd - =

)- c

:h~ ~ ~ C Cd C'3 mtn O S SOa~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

lz

CQX

+3+~

+

0<

x

0q

I0.1

x

+Q

0i

326

Cd 0-S

c0+,

D_E+

>

0 u

Cr-._

S-

. .

r-

_t

0

Cd

0-0r.

- Cd

: 3

0._

Cs

.E OL>

I .CX. o.2 4w

P4r.

-00!

*_

-0

I-

C-

-0

C-C

*_

0

C.

C.)

S

-0

0_

I-0C'sC

+ 0

UV'< CCI)0

+>D0-CC, Od-

_4

+

0_CEix

+

+

0SQq

tCC-

8 tl---- 8 01,00 8 iI" -4 1-4

on Septem


mbr.asm

.org/D

ownloaded from



the comparable strains grown with aeration(data not shown).

Effects of starvation for an energy source priorto and during mating. The data obtained in theexperiments just presented indicated that starva-tion of either broth-grown parent for all carbonand nitrogen sources could cause a reduction inrecombinant yield. To determine more accuratelythe nature of this effect, Hfr and F- cultureswere grown in minimal medium and then starvedfor glucose. The results of such experiments onrecombinant production and zygotic inductionare given in Table 3. It is apparent that the effectson the F- parent of starvation for glucose priorto mating are reversed if glucose is present dur-ing mating. No such restoration of complete fer-tility returns to glucose-starved Hfr parents whenthey are mated in the presence of glucose (Table3). This is probably due to the fact that F pili,which are lost during starvation, must be re-grown before specific pair formation can occurwith an ensuing transfer of genetic material.

The results presented in Table 3 indicate thatactive energy metabolism is probably required inboth parents during conjugation. To reduceenergy metabolism in one parent while allowingit to proceed normally in the other, Hfr and F-mutants unable to utilize maltose were isolated.These strains, which were either able or unableto utilize maltose, were mated in all combina-tions in the presence of either glucose or maltose(Table 4). The results obtained clearly indicatethat maximal frequencies of recombinant pro-duction and zygotic induction are only attainedwhen both the Hfr and F- parents are capableof unrestricted utilization of an energy-yieldingcarbohydrate.

Function of energy metabolism in each parentduring conjugation. Interrupted mating experi-ments with the mutant strains used to collectthe data in Table 4 were never very rewarding.First, the recombinant frequencies obtained inmatings with cultures grown in minimal mediumwere always very low. Second, no plateau in

5 i10250 40 50 20

-0

a- (IO_~~~~*1. Interruped matings betwenHfrandF- arn abl o n /0 0u500 ~~ ~~~~~00 0 30 -1z

Z / 00 0~~~~~~~~~~~

z ~~~~~~~~~~~~~~~~~~~~z /1I0,15 75i /ac75i a.pyr+Utrxy 801) rxyl+ (x800). () Hfr/OR2pyr+ sr o(x503) u<< < 20 A

py) sr y c75.Tebceiweegonuminia medium cotiigaeieUrci,tytpa,hsi

z /

Hf parnt to abu 5 0 el/lwtotarto oaheemaia ubr fFpl e el.Tebc

Z ~~~~~~~~~~zZ

0 0~~~~~~~~~~~~~~~~00) / 0LEr cr~~~~~~r /

5- ~~purE'~ 0--oo-0

were~~~~~~0trefo 30mnadte/yoe()adguoe()wreaddth prpit utrs1 i e

/0-< ~~yr'r o.- /- pyr 0

o - 9 "_ 0 O0

matOFINgsRPTO (mi600 TIME OF INTERRUPTION (min)0a 30Itruteaig between H"Jfr and F- parents able or unable to utilize xylose. (A) Hfjr CR21 purE+

pyr+ stro xyl- Cx 801) X F-purE-pyr- strr xyl+ Cx 800). (B Hfr CR21 purE+pyr+ stre xyl+ (x503) X F-purE-pyre Strr Xytn cx 77). The bacteria were grown in minimal medium containing adenine, uracil, tryptophan, histidine, glucose, and 0.1% Casamino Acids. Theaf parents were grown to about 108 cells/ml with aeration and theHfr parents to about 5 X 108 cells/mi without aeration to achieve maximal numibers ofF phl per cells. The bac-teria were sedimented, washed gently, and resuspended in prewarmed minimal mating medium containing adenine,uracil, tryptophan, and histidine, and the Hfr cultures were adjusted to a cell titer of about 108/mi. The bacteriawere starvedfor 30 min and then xylose (0) and glucose (0) were added to the appropriate cultures 10 mini be-fore miatings commenced. The matings were interrupted by diluting into UV-irradiated T6 (suspended in minimalmedium containing sufficient glucose to give afinal glucose concentration of 0.5%) and immediately agitating witha vortex mixerfor 15 sec. Unadsorbed T6 was neutralized with antiserum to T6. All manipulations were conductedat 37 C.

327VOL. 32, 1968

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL.

recombinant frequency was ever attained. Inaddition, Hfr H and the F- strains derived fromW945 were not closely related. To circumventthese problems, further studies on the functionof energy metabolism during conjugation em-ployed isogenic Hfr and F- parents and madeuse of the fact that almost 100% specific pairformation could be achieved if the donor strainswere grown without aeration in minimal mediumcontaining Casamino Acids (Curtiss and Caro,in preparation).

Figure 1 illustrates the results obtained ininterrupted matings, conducted in the presence ofeither glucose or xylose, between parents eitherable or unable to utilize xylose. Experiments likethose presented in Table 4 were also done withthese strains and similar results were obtained,although the recombinant frequencies were muchhigher than those listed in Table 4. The resultsobtained in interrupted matings between xylose-utilizing parents in the presence of either glucoseor xylose were the same (data not shown). Theresults shown in Fig. IA indicate that supply ofa nonutilizable carbohydrate to the Hfr parentdecreases the number of donor cells capable ofinitiating chromosome transfer without affectingthe rate of chromosome transfer. These inferencesare based on the observations that: (i) the timesof first appearance of purE+ and pyr+ recombi-nants are the same in both matings, (ii) thelength of mating time between the time of firstappearance of either marker and the time ofachieving a plateau frequency for that marker isthe same in both matings, and (iii) the per centinhibition in the frequency of purE+ recombi-nants after 60 min of mating is about equal tothe per cent inhibition in the frequency of pyr+recombinants.The data shown in Fig. 1B indicate that sup-

plying a nonutilizable energy source to the F-parent results in a decrease in the rate of chro-mosome transfer. The reasons for this interpre-tation are that: (i) the time of first appearanceof pyre recombinants is delayed 15 min for themating in the presence of xylose, whereas thetime of first appearance of purE+ recombinantsis the same in both matings; (ii) the length ofmating time between the time of first appearanceofpurE+ recombinants and the time of achievinga plateau frequency of purE+ recombinants ismuch longer in the mating in the presence ofxylose than in the mating in the presence ofglucose; and (iii) the per cent inhibition in thefrequency of recombinants for the distally trans-ferred pyr+ marker is greater after 60 min ofmating than the per cent inhibition in the fre-quency ofpurE+ recombinants.Only some of the interrupted mating experi-

ments conducted to assess the parental require-ments for energy metabolism during conjuga-tion yielded results like those in Fig. 1. There-fore, certain comments about the difficulty ofconducting these types of experiments and thevalidity of the conclusions drawn from them arenecessary. Plateaus in the frequency of recombi-nants are only attained if all of the specific pairswhich can form, are formed within the first fewminutes of mating. This required that essentiallyall Hfr cells have F pili at the inception of star-vation and that we not starve Hfr cultures for agreat length of time prior to mating, whichcould cause a substantial loss of F pili. To ac-complish this, we grew the donor cultures in thepresence of 0.1% (w/v) Casamino Acids to highdensity without aeration and then diluted themafter gentle washing for starvation in mediumwithout Casamino Acids. The continued presenceof Casamino Acids during mating provided theF- parent with enough energy so that there waslittle or no inhibition in recombinant yield or rateof chromosome transfer in experiments like theone presented in Fig. 1B. The total density ofcells in the mating mixture and the ratio ofdonor cells to recipient cells also affected theresults. The inhibitions on either parent weremarkedly reduced if the total cell density duringmating was high (ca. 5 X 108/ml). Also, whenthe Hfr to F- cell ratio was 1:1 to 1:2, thedegrees of inhibition achieved were greater inmatings with the xyl- Hfr and smaller with thexyl F- for matings conducted in the presenceof xylose. Presumably, these effects are due tobreakdown of xylose by the xyl+ parent withutilization of the excreted breakdown productsby the xyl- parent. We also obtained higherrecombination frequencies and better overallresults if the parents were grown on glucoseprior to starvation rather than on glycerol orsuccinate. We demonstrated that the xyl+ strainsused could switch from utilization of glucose toxylose and vice versa with little or no lag.To sum up our experience with the type of

experiment presented in Fig. 1, it can be statedthat we never observed changes in the time offirst appearance of markers donated by Hfrparents unable to utilize the supplied carbonsource but we frequently observed delays in thetimes of appearance of distally transferredmarkers in matings with F- parents unable toutilize the supplied carbon source. Therefore, weconclude that both the Hr and F- parent requirea metabolizable energy source during conjuga-tion, and we suggest that the Hfr and the F-require energy to initate chromosome transferand to control the rate of chromosome transfer,respectively.

328 BACTrERIOL. REV.

on Septem


mbr.asm

.org/D

ownloaded from



0. 0CA,, .o

U8-r-

in t4. trn

*z

0 8

o 5

C'

'5 ~~~~~~~~~~~~1.

.-2

X 4 8o t ~~~88 80

s,

5.,x

~~;-~~~~- ;; ;;0

CCd

A~~~~

~ -xx+ xx xx

0 X XX

0~~~~~~.=d >4o o 4 zo

5.E xx xx xx

00 "' < o ~oo

5.. 5.. 5.. .

In~~~~0 5.. 5..

Cs cq~ . .

k ?<1

c*<d

00

0

en

+6

+t

x

..

+d

0,0

oo

x

+LI...

d0

ioit

VOL. 32, 1968 329

Cs

SkboL.S

.2:

*t!

.t

S:

,I.

Ci

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL.

Effects of Inhibiting DNA Synthesis Before andDuring Conjugation

Recombinant production in matings with thy-mine-requiring parents. Pritchard (48) used thy-Hfr and F- parents to show that starvation ofboth parents for thymine during mating did notalter the rate of chromosome transfer, but diddecrease the total recombinant yield. He alsoobtained this effect on recombinant yield whenonly the Hfr thy- strain was deprived of thymine;Pritchard therefore suggested that DNA synthesismay be required in the Hfr to permit initiationof chromosome transfer. The maximal recombi-nation frequencies in Pritchard's experimentswere only about 5 to 10%. Thus, the averageamount of chromosomal material transferredfrom each donor was low (see 53). It was there-fore possible that the small amount of thyminesynthesized by the possibly leaky, aminopterin-derived thy- Hfr mutant might be sufficient toallow replication of the Hfr chromosome duringtransfer according to the Jacob-Brenner-Cuzin(38) model.We have repeated Pritchard's experiments using

isogenic donor and recipient strains and undermating conditions in which essentially every Hfrcell was paired with one or more F- cells. Table 5contains data from representative matings withaminopterin-derived thy- parents conducted inthe presence and absence of thymidine. Summaryvalues for the mean per cent inhibition in recom-binant yield obtained for each type of matingperformed in the absence of thymidine are alsoincluded. Thymine starvation has little effect onthe thy- F- parent but has a significant effect onthe thy- Hfr parent. The per cent inhibition whenboth thy- parents are deprived of thymine (57%)exceeds the combined inhibition when either thy-strain is singly deprived of thymine. This suggeststhat some supply of thymine or of its derivativesmay be provided to the thy- parent by the thy+parent, either by cross-feeding through the me-dium or by cytoplasmic transfer, or by a com-bination of these methods. The per cent inhibitionin recombinant frequency was usually greater forthe proximally transferred marker than for thedistally transferred marker (Table 5). We have noexplanation for this observation.

Figure 2 presents results from interrupted mat-ings conducted in the presence and absence ofthymidine between thymine-requiring parents.The times of first appearance of the donor thr+and purE+ markers in recombinants are the samein both matings. The times of attaining plateausin recombinant frequencies appear to be slightlydelayed in the mating conducted in the absenceof thymidine, but the beginning of thyminelessdeath of the F- parent, 70 min after removal of

z

0zc)0

z

0

zaLm~ui

-D

/purE °"0O., \

-~~~~At D

20 30 40TIME OF INTERRUPTION (min)

50 60

FIG. 2. Interrupted mating between thymine-requir-ing parents in the presence (solid lines) and absence(dashed lines) of 4 ,ug of thymidine/ml. x 584 (HfrOR41 thr+ purE+ thy- stre) and x 724 (F- thr purE-thr strr) were at titers of 1.2 X 107/ml and 1.9 X108/ml, respectively in the mating mixtures. Othermethods are described in the footnote to Table 5.

thymidine, complicates this type of analysis. Notethat the frequency of recombination for theproximally transferred thr+ marker approaches100% in the mating with thymidine present (seealso Table 5). If the probability of integrating amarker is 0.5 of the transfer frequency (39), theneach Hfr cell in the control experiment presentedin Fig. 2 must be transferring a partial chromo-some to an average of two F- cells.DNA synthesis and initiation of chromosome

transfer. As a first guess as to why thymine starva-tion of thy- Hfr strains inhibited recombinantformation, we decided to test Pritchard's sug-gestion (48) that DNA synthesis was needed inthe Hfr parent to initiate chromosome transfer.To do this, thy- parents were starved for thyminefor 20 min, mated for 5 min in the presence ofthymidine to allow specific pair formation andinitiation of chromosome transfer, and then di-luted 1:200 into medium with or without thymi-dine or thymine. The results (Table 6) indicatethat 5 min in the presence ofthymidine or thymineis sufficient to allow mating partners to giveuninhibited recombinant frequencies during an-other 60 min of mating in the "absence" ofthymine or thymidine. There are at least two types

I I~~~~~~~~~~~~~~~~~~~~~~

330 BACrRIOL. REV.

on Septem


mbr.asm

.org/D

ownloaded from



of potential problems with this type of experi-ment. First, it is possible that the diluted thymine-requiring mating partners can obtain a sufficientamount of thymidine or thymine, which is presentat 0.02 ,ug/ml, from the medium to allow replica-tion of the donor chromosomes during transfer.If uptake of thymidine or thymine were by simplediffusion, then there would be a sufficient amountwithin each cell to allow replication of about 10%of one chromosome per cell if the enzymes in-volved in this replication have high affinities forthymidine or thymine. Ifthe thymidine or thymineis actively transported into the cells, then theconcentration of thymidine or thymine in themedium is sufficient to allow extensive amountsof chromosome replication and cell growth. How-ever, if such uptake occurs, the thymine orthymidine must be preferentially used for replica-tion of the chromosome being transferred and notfor vegetative chromosome replication, since thy-mineless death of both parents is occurring by 80min after dilution of the mating mixture (85 minafter commencement of mating). In control ex-periments, both parents commence thyminelessdeath 70 min after removal of thymine from thegrowth medium (see Fig. 2). It thus seems un-likely that the thymine or thymidine at diluteconcentration is being actively utilized by thethy- parents. The second difficulty concerns thefact that strains in the W1485 subline of E. coliK-12 possess a thymine pool sufficient to allow amean replication of about 10% of one chromo-some per cell (L. G. Caro, personal communica-tion). Since the purE+ marker is about 20 min oftransfer time from the origin of chromosometransfer of Hfr OR41, it would again be necessaryfor the cells to use the thymine pool preferentiallyfor the replication of the chromosome beingtransferred. Furthermore, in one experiment thepyr+ marker, which is transferred 16 min afterthe purE+ marker, was also used for recombinantselection, and the frequencies of pyr+ recombi-nants after 60 min of mating were the same forthe mating partners diluted into medium with andwithout thymidine.The results presented in this section suggest, but

do not prove, (i) that DNA synthesis is requiredin the Hfr during the beginning of mating topermit all the cells to initiate chromosome transferand (ii) that although DNA synthesis may nor-mally accompany chromosome transfer, as recentevidence seems to suggest (6, 33, 34, 49), it is notnecessary for transfer and therefore does notcontrol the rate of chromosome transfer. Ourresults with thymine-requiring strains are also incomplete accord with those obtained by Pritchard(48).

Effects of Inhibiting Protein Synthesis Beforeand During Conjugation

Effects ofamino acid starvation on recombinantproduction. Krisch and Kvetkas (41) and Fisher(29) both showed that amino acid starvation ofamino acid-requiring Hfr strains reduced recom-binant yield. They also showed that the inabilityof Suit et al. (54) to demonstrate an effect ofamino acid starvation on Hfr parents was due tocross-feeding which occurred on the membranefilters used for matings. We have shown thatstarvation of donor strains for a required aminoacid results in an exponential loss in number of Fpill and in specific pair-forming ability, with aconcomitant increase in recipient ability (i.e.,phenocopy production). These results will becommunicated in detail elsewhere (Curtiss andCaro, in preparation), and the data presented anddiscussed in this report principally concern othereffects of amino acid starvation.

Table 7 presents data on the effects of aminoacid starvation of either or both parents onrecombinant production. The most striking fea-ture of these data is the observation that aminoacid starvation of the F- parent before and duringmating results in an increase in recombinantfrequency, especially for the more distally trans-ferred pyr+ marker. Amino acid starvation of theHfr parent causes a decrease in recombinant yieldwith some, but not all, ofthe effect probably beingdue to loss of F pili (see below).

Effects of completion of vegetative chromosomereplication on recombinant production. One well-known effect of amino acid starvation is that itallows completion of a round of vegetative chro-mosome replication without permitting initiationof a new round of replication (42, 46). The deriva-tives of the E. coli K-12 substrain W1485,specifically x 584, behave in this manner (L. G.Caro, personal communication). We therefore per-formed experiments with both Hfr and F- strainsto determine whether there was any correlationbetween recombinant-forming ability and anystage in the chromosome replication cycle.

Figure 3 presents data from an experiment inwhich X 584, a pro- Hfr parent, was starved ofproline to allow completion of vegetative chromo-some replication. The results obtained from mat-ings initiated during the 90-min starvation periodfor proline indicate that amino acid starvationcauses a progressive loss in recombinant-formingability. The addition of proline at the commence-ment of mating only gives a partial restoration ofdonor activity, which is not much greater for theproline-starved Hfr cultures than for the zero-timemating with the unstarved Hfr culture. This lossof donor ability which is restorable by addition

331VOL. 32, 1968

on Septem


mbr.asm

.org/D

ownloaded from


332 CURTISS ET AL. BAcnnuoL. RFv.

c; Q;.61,10 0

X X X X 0WI rA

*.&.4 0eli 1.4 .6 -

0

oo 0 0 0 0 cd

x x x x .5

. E.-,.o 1-4 00 u .54

000 8) a a0 >10 W

X X X X W 0 o-0 0 0 -c;

Cd W! --4

1-40 w

x x x x 0

06 06 +Oa 0co 00 co

0 0

Cd Tl Cd 0x x x x

cli V--4 "-4 4-t >%0=-.4= 0 0

zt +, M4-4 cu0 Qw 0

E 0J. 0x x x x

cd Cd

0tn eq . 4

eq N 0 4-b

0.0 Cdw

Cd ;.., 0W) N 4) CO)tn tn 0 '.. d)

%- 0 - T)

0cdr Cd

0

Cden

Qt 0 C)0x tn .o mt4a P4 4.. 00 0 Cd 4)C; C; C; W) 4) *-b = Wcn

rA .4.00C4

0C4 4;;, 00 0

rAbo

-4 odbo co 'a ..-.

CA

>b0= 0Cd - 0 40- MO

0 'E(L) 04 -0 0

0 C40 C,)

0 Ca (U 00 0 0 0.0 00

JD .0 .0< < 0 O .,.O

0

> O >bcq

4-b *J,

on Septem


mbr.asm

.org/D

ownloaded from



0

0

'I0

0

0

p4

0,

~0 N*- tn4 eq

1..0,

4-.

L..

8 0% 00- tn

-o eni e

80 oo- 1. 00 t-

"_ .

X -- --c -4 - -4

0 + XXX + xxx xxxst mO x tot otS~~a-,tC3~~~ - - , N1 _ _ O% _

0

^

Q 4 xxx txxx xxx~. . - . -- .--..

Pe

bo

,,,-a, = 0C C C

4-4Z E4 Z Z

0

En

C d 0 0 0O a

ZZE ZZZ ZZ;0

bO )a): W V'0.E~ ~ ~ ~ 0 00 0 0 1- w 0 0-

zzz z~Z

p4

+C.

4-,

+ -t

t-. I

oi

en 0

X X

+d

IL1~

ci +

:3

- -.

00 C%IIt xt-x x

-.4-,1

40L.+k

~

t

I..

_L.

IX4 X

U'c.0

> 0

r*b o *3 o

Cd 0

C._C C0

04

4- Cd CA}o

cis

004

.: °°0>

C) eVC)'00 4).YE .:

0 . 4-D

00 CCA

nz (L cd C)

°d 0D °

0,.0

oCA = s

co

CO)3- coe

0

-0

a) C

*d ,d a () 00

La

0

00; *,ad 4)

On °0'0c 0

01-

0 c'0'(L)4-4b00 OL.

Ca 0

- 0 3

*0a 04 0

cd )0)0

4 -. C .) 0

VOL. 32, 1968 333

0

t!CZ

0Q.

0,2-0

0

0

0

ca

0

.s:

0

0

4z

C0

4-,3a)a)

C-

F-

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL.

0~~~~~~~~~~~~~~~~~~~~

o I"

0 0 0 00 ~ 15 0-21 20 27

TIME~~~~~~~~~~~~~fMATNGBEAN(mn

(I)

yildbya Hr arn. 54(Nr R4 hr po purEsfr)adx73(Mtrrour sr)weued

z

0LOi

0 30 60 90 120 150 180 210 240 270TIME MATINGS BEGAN (min)

FiG. 3. Effects of amino acid starvation and completion of vegetative chromosome replication on recombinantyield by an Hfr parent. x 584 (Hfr CR41 thr+ pro- purE+ str') and x 723 (F- thr- pro+ purE- strr) were used.Media andprocedures for growth, starvation, and mating were as described in the footnote to Table 7. The experi-ment was initiated immediately after resuspension of the washed x 584 culture in medium devoid of proline. Aseparate aerated culture ofx 723 was usedfor each mating. Upon addition ofproline to the x 584 culture, periodicdilutions with prewarmed medium were made to maintain the density at about 108 cells/ml. The matings wereinterrupted after 33 min. Recombinant frequency values obtained in matings performed in the presence ofprolineare connected by solid lines and those obtained in the absence ofproline by dashed lines. The points connected bydotted lines are the values for matings in the presence ofproline corrected for loss ofF pili.

of proline to the mating mixture might be due tothe inability of amino acid-starved donors toinitiate chromosome transfer, as was suggestedby Fisher (29). The nonrestorable loss of donorability during amino acid starvation is thereforeprobably due to loss of structural componentsnecessary for conjugation which cannot be suf-ficiently resynthesized within the mating period torestore normal donor activity. F pili, which arenecessary for specific pair formation, are lost atan exponential rate during amino acid starvation(Curtiss and Caro, Bacteriol. Proc., p. 27, 1966;manuscript in preparation), and we have thuscorrected the recombinant frequencies to accountfor the decrease in mean number ofF pili per cell(Fig. 3). Even though no correction was appliedto the data for any decrease in F pili length, itseems reasonable to assume that some othercomponent necessary for conjugation is also lostduring amino acid starvation. Based on studieswith the minicell-producing mutant isolated anddescribed by Adler et al. (3), Cohen et al. (14,16) inferred that donor and recipient cells differwith respect to a cell surface component inaddition to the presence or absence of F pili.Although it was suggested that this cell surface

component was involved in pair formation, theexperiments did not permit the distinction as towhether the cell surface component(s) was on thedonor or the recipient cell, or on both. We(Curtiss and Stallions, unpublished data) haverecently isolated two classes of mutant donorstrains which have relevance to this discussion,one which has F pili but is unable to donategenetic material and the other which lacks F phlibut has the recipient ability characteristic of theparental donor strain. Mutants of the first typecan form specific pairs with F- cells, whereasmutants of the second type cannot. Mutants ofthe second type form specific pairs with donorcells infrequently. On the basis of all of theseobservations, we suggest that much of the loss ofdonor ability during amino acid starvation, whichis not restored by addition of proline to themating and which is not accounted for by F pililoss, is probably due to loss of a donor cellsurface component necessary for effective pairformation.

If ability of Hfr donor cells to initiate chromo-some transfer were correlated with some stage inthe vegetative chromosome replication cycle, as iscalled for in the Bouck-Adelberg (8) model for

334 BACTERIL. REV.

on Septem


mbr.asm

.org/D

ownloaded from



80 00

-a_zI

o I I

0~~~~~~~~~~~~~~~~

° 50 I

z I

o I /

w 40

V) 0~~~~0

00-

0~~~~~~~~-

20 30 40 50 60

TIME OF INTERRUPTION (min

FiG. 4. Effieci's of amino acid starvation and com-

pletion of vegetative chromosome replication on re-

combinant yield by an F- parent. x 503 (Hfr 0R21

thr+ purE+ pyr+ hisP strt) and x 820 (F- thr purE-pyr~ hisn strr) were used. Media and procedures for

growth, starvation, and mating were as described in

the footnote to Table 7. x 820 was starved (solid lines)

or not starved (dashed lines) for histidine and threonine

,for 60 mill prior to mating. Histidine and threonine

were added to the starved x 820 clultuare 30 sec before

addition of the H,fr parent.

chromosome transfer, then it might he expected

that recombinant frequencies would fluctuate in

some cyclic manner after readdition of proline to

the Hfr culture. This type of result was not

observed, and the recombinant frequencies ob-

tained from matings initiated after proline was

restored to the culture do not vary significantly

from those obtained in the mating of the un-

treated controls (Fig. 3). There was also no

observed change in the kinetics of recombinant

formation associated with completion of vegeta-

tive chromosome repLication as measured in in-

terrupted mating experiments (data not shown).

Synchronous reinitiation of DNA synthesis upon

addition of proline to the Hfr culture at 90 minmay be poor. However, the same type of experi-

ment as presented in Fig. 3 has been done with

the F+ strain from which X 584 was isolated, and

the results indicate a correlation between donor

ability and some stage in the vegetative chromo-

some replication cycle (23). The results presented

in Fig. 3 do not, therefore, favor any model which

requires a correlation between a specific stage inthe vegetative chromosome replication cycle andability to initiate chromosome transfer by Hfrdonors.

Figure 4 presents results of an interruptedmating experiment with an amino acid-requiringF- strain which either has been starved of aminoacids for 60 min or has not been starved. Aminoacid starvation has dramatically altered the ap-parent kinetics of chromosome transfer. Thelength of mating time between the time of firstappearance of each Hfr marker in recombinantsand the time a plateau in the frequency of thatrecombinant type is attained has been cut inhalf in the mating with the amino acid-starvedF-. The times of first appearance of each Hfrmarker in recombinants are unchanged by usingan amino acid-starved F-, and we thereforeconclude that there is no change in the rate ofchromosome transfer. The experiment presentedin Fig. 4 provides an explanation for the data inTable 7 which showed that amino acid starvationof an amino acid-requiring F- caused increasedrecombinant yields, especially for Hfr markerslocated distally from the origin of chromosometransfer. The matings used to collect the datafor Table 7 were interrupted after 40 min. It canbe seen in Fig. 4 that the frequency of pyr+recombinants is still increasing at 40 min in themating with the unstarved F-, while it has alreadyreached a plateau in the mating with the aminoacid-starved recipient.

There are at least two possible explanations forthe results obtained with amino acid-starved F-recipients. In the first, it is suggested that vegeta-tive chromosome replication must reach a specificstage (e.g., completion) before the F- can activelyparticipate in the transfer of the donor chromo-some with an expenditure of energy. In thesecond, it is suggested that amino acid starvationalters the physiology of the recipient so thatmarkers which are nearest the end of the chromo-some broken at the time of interruption are notlost by nuclease degradation. A corollary of thissecond suggestion is that, under normal matingconditions, the probability of integrating a markertransferred into the recipient just before the timeof interruption is much less than 0.5. This secondpossible explanation was kindly provided byCharles Brinton (personal communication). Theuse of zygotic induction did not allow a distinctionbetween these two hypotheses, but other experi-ments are being conducted to prove or disprovethe first explanation.

In summary, amino acid starvation of Hfrdonors causes (i) a loss of F pili and ability toform specific pairs with F- cells (Curtiss and

VOL. 32, 1968 335

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL. BACTERIOL. RE\

0.8'10.21* I-a 12' 1 6.5' -2.0 4-2.+2025

accerg leu proA proB ,proC T6 purE gal

proA-8B

proS- /oc

/ I/ \

F STRAIN Hfr H Hfr ORII Hfr OR6Hfr P4X6 HfrORI Hfr OR7 Hfr Cav/eut prof //I, po,f 1e1,t pro A*1 pro* A/e1 pro" pro* leou pro*

-100 i

eupro|InI[1 A GS75X680

Eli [I ~~~~~~~~~~50leu proAl4El1

25 Z

5~~~ 0

X137

/eu -proA -B8 -

100

-75

-50

- 25

-0

uJ

LLJ0-j

X277

/eu proc |

F1-1. 5. 1/IC pro-oCIc regioli oJ the E. coli chromosome, the origilns anid cdirectionis of chIromosomi7e tri-l(s/cr h)lHflr donors, and the effect of a proA -B- culetiont mutation on recombinanlt yield. The F- strains x 680 (proA-),x 278 (proB-), x 277 (proC ), and x 137 (proA-B-) are all closely related andl are descended from the C600 sub-line. Hlfr H (x 57), Hf'r P4 X 6 (x 313) and Hf'r Cav (x 235) are all descended.from the 58-161 subli/me. Hfr OR](x 225) and Hfr ORII (x 493) were isolated from the same F+ strain and are descended f'rom W1485. Hf'rOR6(x 435) and IHfr OR7 (x 436) ar-e closely related and are descended from the K-12-112 subline. Tlhe data presenitedwere obtained by mating each Hflr strain withl the four F- strains on two to five different occasionts. Thle average

recombinant frequencies were used to calculate relative values. (The frequency of leu+ recombinants in crosses.

wvith x 680 or ofproC+ recombinants in crosses with x 277 wvas equated to 100.) Bacteria were grown) and mitatedlin L broth. Matintgs were of'40-mnm dlurationi.

Caro, Bacteriol. Proc., p. 27, 1966; manuscript inpreparation), (ii) a possible loss of a structuralcomponent of the donor cell surface which is

needed to establish effective pairs with F- cellsand prevents donor-donor matings, and (iii) a

possible inability to synthesize a specific proteinneeded by the Hfr to initiate chromosome transfer(29). Amino acid starvation of the F- recipientalters the apparent kinetics of chromosome trans-fer by either of two presently indistinguishablemechanisms.

Effects of Genetic Inhomology Between Donorand Recipient Chlromosomes on

Chlromosome Transfer-

In 1964, Johnson, Falkow, and Baron (40)demonstrated in matings between Hfr strains ofE. coli K-12 and recipient strains of Salimonellatyphosa that large increases in recombinant fie-quency were obtained when the recipient was a

hybrid containing E. coli genetic material. How-ever, these increased recombinant frequencieswere only observed when the hybrid recipient

336

on Septem


mbr.asm

.org/D

ownloaded from


CONJUGATION IN E.COL3

contained genetic material from the proximallytransferred portion of the Hfr chromosome, butnot when such hybrids contained distally trans-ferred E. coli genetic material. We (20) haveobserved a similar requirement for genetic homol-ogy between the lead region of the Hfr chromo-some and the comparable portion of the F-chromosome in E. coli K-12.

Effects of deletions in the F` chromosome onrecombinant production by Hfr donors. The di-agram at the top of Fig. 5 provides a map of theportion of the E. coli chromosome used for thesestudies, with the origins and directions of chromo-some transfer for seven Hfr strains. Our initialobservation of an effect of genetic inhomology onrecombinant yield was obtained in matings inwhich F- strains possessed the proA-B deletionmutation. This mutation deletes a segment ofchromosome which includes the proA and proBcistrons and, in addition to causing a requirementfor proline, confers resistance to four phages (18).Several independently isolated proA-B- muta-tions have been intensively studied, and resultsfrom a variety of genetic tests have indicated thatall have the same termination point between theproB and lac loci (Curtiss and Charamella, un-published data). However, the termination pointbetween the argF and proA loci (see 55) has notbeen determined, and therefore the segment de-leted in proA-B- strains is between 2.2 and 4.5%of the chromosome.The data presented in Fig. 5 summarize the

results of matings between seven Hfr strains andF- strains having either a point mutation at theproA, proB, or proC locus or a proA-B deletionmutation. All four F- strains had the same leu-mutation. In the matings with the proA- F- (X680), all Hfr donors gave results which weresimilar and reasonably normal. The only dis-crepancy from expectation was the fact that thefrequency ofproA+ recombinants did not exceedthe frequency of leu+ recombinants in the matingswith Hfr ORli, Hfr OR6, Hfr P4X6, and HfrORL. This is a property of all proA- derivativesof the C600 subline and is not found with proA-mutants of the W945 F- subline. In the matingswith the proBT F- (x 278), there is a significantreduction in the frequency ofproB+ recombinantsin matings with Hfr ORi 1, Hfr OR6, and HfrP4X6. This is probably due to the proximity ofthe proB locus to the origins of chromosometransfer for these Hfr strains (32, 45, 47, 58).The frequencies of leu+ recombinants are, how-ever, very similar to the frequencies of leu+recombinants found in matings with x 680 andx 277. In the mating in which the F- has theproA-Bc deletion mutation (X 137), some of theresults differ significantly from those obtained in

matings with F- strains having point mutations.The frequency of leu+ recombinants is normal inmatings with Hfr H and Hfr Cav, and is reducedby a factor of two in the mating with Hfr OR7.However, the frequency of leu+ recombinants isreduced 24-, 18-, 14-, and 34-fold in matingswith Hfr ORil, Hfr OR6, Hfr P4X6, and HfrOR1, respectively. Thus, a significant reductionin the frequency of leu+ recombinants is onlyachieved when there is a large deletion in the F-chromosome for the segment of chromosome firsttransferred by an Hfr donor.

Similar results leading to the same conclusionwere obtained from matings between the Hfrdonors used to obtain the data in Fig. 5 and aseries of proline-deficient F- strains derived fromthe W945 subline of E. coli K-12 (data notshown).To obtain more information about the require-

ment for genetic homology between the leadregion of the Hfr chromosome and the com-parable portion of the F- chromosome, weinitiated studies with isogenic strains derived fromthe W1485 subline. Hfr OR1, which has Fintegrated into the right end (Fig. 5) of thestructural gene for ,B-galactosidase (lacZ; Curtiss,Bacteriol. Proc., p. 30, 1964), was used as thedonor. A series of F- strains was prepared havingthe same arg- and leu- alleles but with differentproA-, proBc, and lac- point mutations and withproA-B-, proB-lac-, and lacZ-Il deletion muta-tions (see Table 1). The results from representa-tive interrupted matings between Hfr ORI andthese F- strains are presented in Fig. 6. Re-combinant frequencies for each selected markerare lowest when the F- possesses a proA-B-deletion mutation, are significantly reduced whenthe F- possesses a proB-lac- deletion mutation,and are essentially unaffected when the F- hasthe lacZ-I- deletion mutation.When the data from matings with F- strains

having deletions are plotted with nearly the sameordinate scale used for plotting data from thecontrol matings, there is an apparent long delayin the time of first appearance of each donormarker in recombinants with respect to markerentry times in the control matings (see Fig. 6Aand C). However, when these data are replottedwith the use of an expanded ordinate scale (Fig.6B), it can be seen that the delay in markerentry time is much less, although still highlysignificant. (Note that all of these matings weredone with a donor-to-recipient cell ratio of be-tween 1:2 and 1: 3 to facilitate detection of earlyformation of rare recombinants.)The data presented in Fig. 6 and obtained from

other matings with the W1485-derived F- strainsare summarized in Table 8. First, by comparing

337VOL. 32, 1968

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL. BACTERIOL. RI-\.

18 -

16 -

14 -

12 -

< 10-

_U)

';- 8 -

6

, c~

0 c

/

pro4A

025

/eu -c

0/1,6

LU020

uU)

0Lr 015

010

,C

005

0

~~~~~~arq

LU

2 - 0X ' 0o, proA [050<

0 _-/eu 0258016o- aro

Ic

16 oro,4'

14 -

12 -

LU-lJcr 10-

- 8 -

6 -

4 -

2

LUJ

U)(n)

x

/eu

proA -B#

/eu

/eu *

5- D4 -

2 -

O .. .-

14 - E

12 -

10 -

LU

8 -

U')

6-x

4 -

c' t^ ~~orgq/

o o- - oproA -Bf°' ' O l*-~/eu#

10 20 30 40 50 60

TIME OF INTERRUPTION

2 -

0(mi0

(min)10

110 GO_q

20 30 40 50 6Cl

FIG. 6. Results oJ interrupted matings between Hfr ORI and F- strainis having point or deletionz muttationis illthe region of chromosomefirst transferred by Hfr OR]. (A) Hfr OR] (x 225) X x 710 (proA- leu- arg ) (dashedllines and left-hand ordinate scale) and Hfr OR] X x 711 (proA-B- leu- arg-) (solid lines and right hand X 4 e.x-panided scale). (B) Hfr OR] X x 711 data shown in Fig. 6A using a X 50 expanded ordinate scale. (C) Aniotheerset of matinigs between Hfr OR] and x 710 (dashed lines) and Hfr OR] and x 711 (solid lines). (D) Hfr OR] Xx 784 (proB-lac- leu- arg-). (E) Hfr OR] X x 733 (lac Z-I1 leat- arg-). Bacteria were grown and mated int Lbroth. The donor-to-recipienit cell ratios were betweeni 1:2 an?d 1:3 in the matinig mixtures.

the mean recombinant frequencies for matingswith X 710 with those from matings with X 711,it is observed that the pro+, leu+, and arg+

recombinant frequencies are depressed 82, 95, and

255 times, respectively, when the F- possesses theproA-Th deletion. The normal marker entry timesfor pro+, leu+, and arg+ are also delayed 7.3, 6.0,and 8.0 min. resDectively, if the F- recipient has

338

a-z

z0c

zLL

Y-)LUJa-Uf)

Cr)

z

2

0LLJ0E

20]A

.

.

I

II

on Septem


mbr.asm

.org/D

ownloaded from



+?

C"

C.

r0edr-

u]

x

I I I

-1

)0tn 01_t

xxx0 0 (D Ch

000- -'

C.4 0

. .-

0e CNe6

I I

00 0".400

0 0 tn tri r mtn N -4 -"t tn It llT C1 C1 N eq C1 C1

4) ---- --

- - eW) et en-- - -

+IP

"0

0C)

C'

-4

x

.-I

xxx0o es4 W tn

r" W-4 V--

r-

I I

II II

10U) U0

wo~~~~~ees

00.0o + O ' 0 l'.oet

-SOU) o 1-4 -4oI

W-)000 0 0I-RCi%RCi f! (0

0

04.toeo

Ca

"0

'4

-2

4."

4.)

0s

£C.0L)0)

,

ed

,x

0.CO00

0o00)

0Q0.o0)

0)

0

CO

C)

0C0

~ _

.0 0 '

- o-~

".0 E~d 0)CC-O

0)

"0o00)

Y 040n:"v.

-COCOs

e 0o0)c"

VOL. 32, 1968 339

1-

c-

. -

i:r_

bo

C4..

0s-"0

"t

*, I

"0

4

ts; e06

"0

0 S*0)CC0)oo

4..

_0

0-"

-"

.4 O

4.S

4.

-"0

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS Er AL. BACTERIOL. REV.

a

a-

za:

0z00

z0

ujC)

z

zLuJ

a-(f

a:

0 10 20 30 40TIME OF INTERRUPTION (mmn)

FIG. 7. Interrupted matings to select leu+ strr recombinants using four Hfr OR1 strains having independentlyisolatedproA-B deletion mutations (closed symbols and left ordinate) and the Hfr OR1 parent (x 225; open circlesand right ordinate). x 711 (F- proA-B- leu- strr) was used as the recipient and was at a titer of about 2 X 108/mlin the mating mixtures. The Hfr titers were 107 to 2 X 107/ml in the mating mixtures.

the proA-Th deletion. The data from matingswith X 784, which has a proB-lac- deletion,indicate that leu+ and arg+ recombinant fre-quencies are reduced 3.4 and 5.7 times, respec-tively, and that leu+ and arg+ entry times areeach delayed by 2.3 min. The lacZ-I- deletion inX 733 exerts no significant effect either on re-combinant frequencies or on marker entry times(Table 8).

Effects of deletions in the Hfr chromosome on

recombinant production. To eliminate any spuriousexplanations for the effects noted above, it isnecessary to show that "normal" behavior isrestored in matings between Hfr and F- strainshaving the same deletion mutation. Unfortu-nately, the proB-lac- deletion spans the integratedF in Hfr ORI and the proA-B- deletion mutationconfers resistance to the transducing phage Plkc.Therefore, it was not possible to introduce any ofthe identical deletion mutations from the F-strains into Hfr ORI. However, the proA-B-deletions are probably ditto deletions with similar,if not identical, ends, as has already been men-tioned. Therefore, several independent proA-B7deletion mutants of Hfr OR1 were selected bychallenge with phage T7 and replica plating toproline-deficient medium (18). When these HfrOR1 proA-BC donors are mated with X 710, anF- with a proA- point mutation, normal recom-binant frequencies and entry times are obtainedfor leu+ and arg+ (data not shown). Figure 7presents data from matings between four of theseHfr ORI proA-B- mutants and the proA-B F-

X 711, and also data from a mating between theoriginal Hfr OR1 and x 711. The introduction ofthe proA-B- deletion into the Hfr restored "nor-mal" frequencies of inheritance and entry timesfor the leu+ marker.

All of the results presented in this and thepreceding section indicate that genetic homologybetween the lead region of the Hfr chromosomeand the comparable section of the F- chromo-some is required for early appearance of markersin recombinants and maximal frequencies ofrecombinant formation. There are two possibleinterpretations of these results. First, a lack ofgenetic homology between the lead region of theHfr chromosome and the comparable segment ofthe F- chromosome could have no effect onchromosome transfer but prevent the associationbetween a normally transferred donor chromo-some and the recipient chromosome, with aresultant loss of recombinants. Pittard andWalker (47) have obtained data which indicatethat an interaction between a region near theleading end of the donor chromosome and thecomparable segment of the recipient chromosomeis associated with recombinant production. Sec-ond, it is possible that effective homologouspairing between the lead region of the Hfrchromosome and the comparable segment of theF- chromosome must be established before an

appreciable amount of chromosome transfer canoccur. These two hypotheses are not mutuallyexclusive and together may account for theresults obtained. A means of measuring chromo-

340

-_l

DQ-

z

.5 ir

z0

z

ur.0w

z

o2

0

0u(i)

U)

50 60

on Septem


mbr.asm

.org/D

ownloaded from


CONJUGATION IN E.COL3

A .

80- 80-

z /0 60 60

U /cc

200 40

O220 0 2023040 0 1C 20 30 40

7 ME CF INTERRUPTION (mrn)

FiG. 8. F' lac+ transfer to F- strains with a lac-point mutation (x 708, A) or a proB-lac deletion mu-

tation (x 545, B). x 314 was used as the F' donor andlac+ strr recombinants were selected. Bacteria weregrown and mated in L broth. The donor to recipientcell ratio was about 1:20. All (100 of 100) of the lac+strr recombinants in both matings were partially di-ploid F' lac+ lac- donors.

some transfer which is independent of recombi-nant production must be used to distinguishbetween these two models.

Effects ofdeletions in the F- chromosome on F'transfer. The occurrence of haploid F' strains (5,51) indicates that the replication of and expres-sion of markers on F' episomes does not requirethe presence of the segment of chromosomehomologous to the F' episome. Thus, expressionof F' markers after conjugal transfer should beunaffected by a deletion in the recipient chromo-some for part or all of the segment homologousto the genetic content of the F'. Therefore,studies on F' transfer to recipients with and with-out deletions corresponding to regions of the F'episome afford a means to determine whetherhomologous pairing between the lead region ofthe donor chromosome and the comparablesegment of the recipient chromosome is neededfor integration or for transfer of donor geneticmaterial.The kinetics and frequencies of transfer of the

F' lac+ episome, isolated by Jacob and Adelberg(37), to F- recipients having either a lac- pointmutation or a proB-lac- deletion mutation areshown in Fig. 8. (All of the chromosomal materialpresent in the F' lac+ episome is missing from thechromosome of F- strains with the proB-lacedeletion.) As is shown in Fig. 8, the frequenciesand rates of transfer of the F' lac+ episome toboth recipients are indistinguishable. Thus, ho-mologous pairing between the lac regions of the

F' and chromosome is not necessary for theefficient transfer of the F' lac+ episome. Thetransfer of the F episome also must not dependon homologous pairing between F and the chro-mosome. The observed low probability of Fintegration per bacterium per generation (ca.3 X 10-6; Curtiss and Stallions, Bacteriol. Proc.,p. 55, 1968) suggests that only a limited homologyexists between F and the chromosome, and yet Fis rapidly transferred to F- recipients duringconjugation with F+ donors.The kinetics and frequencies of transfer of the

F' lac+ proB+ proA+ episome ORF-206 (5) to F-recipients having either a proA- point mutationor a proA-Hc deletion mutation are shown in Fig.9. The proA-Hc mutation deletes the segment ofchromosome first transferred by F' ORF-206 andcauses about a 100-fold reduction in F' transfer.Thus, it appears that effective homologous pairingis required for the transfer of the F' ORF-206episome. As was stated in the discussion of resultsobtained from Hfr matings with F- strains withthe proA-Bc deletion mutation, the proA-Bc mu-tation does not appreciably affect recombinantyield in matings with Hfr donors which possessorigins of chromosome transfer distal from thelac to proA segment. To eliminate the remotepossibility that the proA-Bc deletion mutationcauses some type of restriction and loss of F'episomes, matings were performed between adonor with the KLF-1 episome (thr+ leu+ F)isolated by B. Low and the F- strains used inthese studies. The data presented in Table 9show that the frequencies of leu+ recombinantsare essentially the same for F- strains havingproA- and lac- point mutations and proB-laceand proA-Bc deletion mutations.Table 10 presents additional data on F' and

chromosome transfer by F' lac+ and F' lac+proB+proA+ donors to the same four F- strains used toobtain the results presented in Table 9. F' lac+transfer is unaffected by deletion of the proB-lacsegment (mating 3 versus mating 1, Table 10),whereas the frequency of recombinants whicharise as a consequence of chromosome mobiliza-tion and transfer by the F' lac+ episome is slightlydepressed when the F- has the proB-lac- deletionand is greatly depressed when the F- has theproA-B- deletion (matings 3 and 4, respectively,Table 10). In fact, the depressions in the fre-quencies of pro+, leu+, and arg+ recombinants inmatings with the F- strains X 784 and X 711 arevery similar to those observed in the matingswith Hfr ORI (Fig. 6, Table 8). Transfer of theF' lac+ proB+ proA+ factor is reduced almost100-fold in the mating with the F- having theproA-B deletion (mating 8, Table 10), but isunaffected when the F- possesses the proB-lac

341VOL. 32, 1968

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL BACTERIOL. REV.

180-

160-

z14 (v

0

z0

c 120-z

w 100-

U)

H80-z

z

60-

0

40-

201

0 -.- 0-0-0 10

2.4-

.

1.8-

1.5-

-i

U) i2<,z 1.2-0co

0.9-

S,5

B

proA-8' L

-0-O-o ?-oo o0 o -o_o_Oo:o20 30 40 0 10

TIME OF INTERRUPTION (min)

FIG. 9. F' lac+ proBe proA+ transfer to F- strains with either a proA- point mutation (x 710) or a proA-Bdeletion mutation (x 711). (A) x 646 (F' ORF-206) X x 710 with selection ofproA+ strr recombinants (0) andx 646 X xL711 with selection ofproA-B+ strr recombinants (0). (B). Data from the x 646 X x 711 matingplotted with a X 80 ordinate scale. The bacteria were grown and mated in L broth. The donor-to-recipient cellratio was about 1:10. Five per cent of the proA+ recombinants in the x 646 X x 710 mating were haploid andnondonors. All othler pro+ recombinants in both matings were partially diploid donors of F' ORF-206.

TABLE 9. F' leu+ transfer to leu F- strains witheither deletion or point mutations in the lac to

proA region of their chromosomes"

InheritanceMating of leu+

allele

x 927 F' thr+ leu+/thr- leu- X(1) x 708 F- lacZ- proB- leu- arg-. 73(2) x 710 F- proA- leu- arg ......... 56(3) x 784 F- proB-lac- leui arg:....... 115(4) x 711 F- proA-B7 leui arg ....... 43

aThe bacteria were grown and mated in Lbroth. The matings were interrupted after 40 minby diluting the mating mixtures into minimalmedium containing glucose, streptomycin, and10%o L broth (v/v) and immediately agitating thediluted mixtures for 60 sec on a vortex mixer.Further dilutions and plating were done afteranother 15 min at 37 C. The titers of x 927, x 708,x 710, x 784, and x 711 in the mating mixtures were8.7 X 106, 1.4 X 108, 1.3 X 108, 2.8 X 108, and8.4 X 107/ml, respectively.

deletion (mating 7, Table 10). The proA-B dele-tion mutation deletes the segment of chromosomecomparable to the leading end of the F' lac+proB+ proA+ episome, whereas the proB-lac dele-tion deletes the segment comparable to the distally

transferred end of the episome. Thus, effectivehomologous pairing between the proximallytransferred end of the F' lnc+ proB+ proA+episome and the comparable segment of therecipient chromosome can occur when the F-possesses the proB-lac deletion, but not when theF- possesses the proA-B deletion. The F' lac+proB+ proA+ episome contains an inverted lacoperon, and therefore this F' can cause chromo-some mobilization and transfer in both clockwiseand counterclockwise directions (5). The data inTable 10 reveal that chromosome mobilizationand transfer in the counterclockwise direction arereduced when the F- possesses either the proB-lacor proA-B deletion mutation (matings 7 and 8,respectively). Also, the inversion within the F'lac+ proB+ proA+ episome reduces the frequencyof leu+ and arg+ recombinants which arise as aconsequence of chromosome mobilization, evenwhen the F- recipients have no deletion mutations(matings 1 and 2 versus matings 5 and 6, Table10). Based on the results presented in Fig. 8 and9 and in Tables 9 and 10, we conclude that ef-fective homologous pairing between the recipientchromosome and episome is not necessary for Fand F' lac+ transfer, but is necessary for thetransfer of the F' lac+ proB+ proA+ episome andfor chromosome mobilization and transfer medi-ated by F and by F' episomes.

342

2001 A .

2.1

on Septem


mbr.asm

.org/D

ownloaded from



I..

0cu

@_ =

> 0

.0

fn 00 C','cnoo o

C; C; 60000.

0000O

- Cf 000o-.oo0 0 0; Co o Co

\- in m o

O o o 0000

00--400 en-000§0000

v

0'

.~ ~ ~ ~ ~C-.,0 -P4C 0

4)-~~~~~~~~~.

Cs ~ ~ ~ ~

to

..

_

_X. ;; , M.I

t , I

N + N

-> C, a .0 Ct C.0 N0 00t;00 0 i m50 0 __o_ oo_ _ O _00_

q >'m ~t tn vo r-

343

00

N

00

o00

._

(L c,4

Nx

5. 00

U o) +

0~

C4)

r+

CL)

cOr

0.

C3 coU

Co CZ

U Us

VOL. 32, 1968

I-

0l

s..

0,

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL.

The major difference between F and F' lac+on the one hand and the F' lac+ proB+ proA+episome on the other is their size. By using geneticdata and the molecular weights of F (31) and ofthe chromosome (13), we calculate that F, F'lac+, and F' lac+ proB+ proA+ contain 1.7, 2.8,and 5.5%, as much DNA as the chromosome,respectively. If effective homologous pairing be-tween the lead region of a donor chromosomeand the comparable portion of the recipientchromosome is required to achieve a substantialfrequency of transfer, then it follows that amechanism must exist to permit the transfer ofthe lead region of the donor chromosome into therecipient cell. We therefore suggest that all typesof donors, be they F+, F', or Hfr, are capable offorcibly transferring an average of about 3% oftheir genetic material into F- recipients withoutthe need for effective homologous pairing. Furthergenetic transfer, however, would require effectivehomologous pairing. Thus, the lack of genetichomology between episome and chromosomewould not alter the kinetics of F and F' lac+transfer, since these elements are small enough tobe forcibly transferred in toto by the donor cells.For genetic transfer by Hfr donors (and donors oflong F' factors), it is probable that the amount ofdonor chromosome forcibly transferred by eachdonor cell would vary and that the mean amountof material transferred in this manner wouldincrease in proportion to the length of matingtime. In the experiments presented with Hfr ORI(Fig. 6A, B, and C) and with F' ORF-206 (Fig.9), about 10% of the donor cells were able totransfer a sufficient amount of their chromosomesso that effective homologous pairing with thesegment beyond the proA-B- deletion couldoccur.A corollary prediction of the above hypothesis

on F' transfer is that there should be an ob-servable difference in recombinant types issuingfrom matings between F- strains and donors witheither short or long F' episomes. If short F'episomes are pushed into the recipient withoutthe need for effective homologous pairing, thenmost recombinants (ca. 99%) should be heterozy-gous and partially diploid. This is precisely theresult found for the recombinants formed in themating with the F' lac+ donor described in Fig.8A. If the lead region of long F' episomes mustundergo effective homologous pairing with thecomparable segment ofthe recipient chromosome,then some recombinants (ca. 20 to 50%) for leadmarkers from F' will be haploid, and some re-combinants (ca. 5 to 20%) having received theentire F' will be homozygous at some loci andpartially diploid. Results of this type have beenobtained in matings with F' ORF-1, which has

an F' containing 8% of the chromosome (5;unpublished data). The presence of the invertedlac operon in F' ORF-206 probably inhibitshaploid recombinant formation for proximallytransferred episome markers, although the fre-quency of haploid recombinants (5%, see legendto Fig. 9) is intermediate between the frequenciesfound in matings with donors of F' lac+ andF' ORF-1.On the basis of the results obtained in the

experiments described in this section, we con-cluded that: (i) all donor strains, be they F+. F',or Hfr, are capable of forcibly transferring severalper cent of their genetic material into F- recipientswithout a requirement for effective homologouspairing, and (ii) further transfer of the donorgenome requires that the lead region of the donorchromosome forcibly transferred be effectivelypaired with the comparable region of the recipientchromosome. In view of the results obtained byus and by Pittard and Walker (47), we alsobelieve that the effective homologous pairingrequired for chromosome transfer is also inti-mately involved in the first steps leading tohaploid recombinant production.

SUMMARY DISCUSSION

In the first section of the RESULTS, the experi-ments conducted led us to conclude that both theHfr and F- parent require a metabolizable energysource during conjugation and to suggest that theHfr parent requires energy to initiate chromosometransfer and that the F- parent requires energyto control the rate of chromosome transfer.

In the second section of the RESULTS, the dataobtained led us to suggest: (i) that DNA synthesisin the Hfr parent is required at the beginning ofmating to initiate chromosome transfer and (ii)that, although continual DNA synthesis maynormally accompany chromosome transfer, it isnot necessary for transfer and therefore does notcontrol the rate of chromosome transfer.The experimental data in the third section of

the RESULTS indicated that amino acid starvationof the Hfr parent led to a loss of donor ability,most of which was not immediately recoveredwhen the amino acid was restored to the culture.This nonrecoverable loss in donor ability wasinterpreted as being due to loss of F pili (Curtissand Caro; Bacteriol. Proc., p. 27, 1966; manuscriptin preparation) with a resultant loss in specificpair-forming ability and to the loss of a postulateddonor cell surface component needed to establisheffective pairs with F- cells. The restorable loss ofdonor ability during amino acid starvation wasinterpreted to indicate a requirement for proteinsynthesis in the Hfr at the inception of mating for

344 BACTERIOL. REV.

on Septem


mbr.asm

.org/D

ownloaded from



initiation of chromosome transfer (29). Aminoacid starvation of the F- recipient, which permitscompletion of vegetative chromosome replication(42, 46), alters the apparent kinetics of chromo-some transfer; this observation was interpreted tomean either that degradation of the donor chro-mosomal region last transferred at the time ofmating interruption is not degraded in aminoacid-starved F- cells or that vegetative chromo-some replication must reach a specific stage (e.g.,completion) before the F- can actively participatein the transfer of the donor chromosome with anexpenditure of energy. The latter interpretationalso implies that vegetative replication of therecipient chromosome and chromosome transfermay be mutually exclusive processes.

In the last section of the RESULTS, the dataobtained led us to conclude (i) that F+, F', andHfr donors are capable of forcibly transferringseveral per cent of their genetic material into F-recipients without a requirement for effectivehomologous pairing between the donor materialtransferred and the comparable segment of therecipient chromosome, and (ii) that more ex-tensive transfer of the donor genome first requiresthat the forcibly transferred lead region of thedonor chromosome undergo effective homologouspairing with the comparable region of the re-cipient chromosome. We further suggested thatthe effective homologous pairing required forchromosome transfer is also intimately associatedwith the initial events leading to haploid recom-binant formation.The conclusions and suggested interpretations

of the data presented in this manuscript, alongwith the results of studies conducted in otherlaboratories, can be used to construct a con-ceptual model for the steps during bacterialconjugation between Hfr and F- bacteria. Withregard to specific pair formation, the work ofBrinton and associates (9-11) and our work(Curtiss and Caro, Bacteriol. Proc., p. 27, 1966;manuscript in preparation) have shown an obligaterequirement for the presence of F pili on thedonor cell surface. We (22) have also shown thatspecific pair formation occurs in the absence ofactive energy metabolism in both parents. Theconversion of a specific pair, which is experi-mentally defined as a union stable to dilution(26) to an effective pair, which is operationallydefined as a union ready to initiate chromosometransfer, is presently a poorly understood process.We have suggested, from results obtained inmatings with the DNA-less minicells (14, 16),from the isolation and characterization of mutantdonor strains (Curtiss and Stallions, unpublisheddata), and from our results in experiments onamino acid starvation of donor strains, that

donor cells may possess a surface component inaddition to F pili which is necessary for promotingeffective unions between donor and recipient cellsand for preventing matings between donor cells.The establishment of effective cell contact be-tween a donor and recipient cell would then actas a stimulus to the donor cell to initiate chromo-some mobilization, as suggested by Jacob,Brenner, and Cuzin (38).Chromosome mobilization, which is opera-

tionally defined as the production of a donorchromosome being ready for linear sequentialtransfer to a recipient cell, would occur in the Hfrcell and be under the genetic control of F (38).There would be an initial requirement for proteinsynthesis (29, 50; third section of RESULTS) toallow for the synthesis of an F-coded initiator(24, 38) followed by an obligate requirement forDNA synthesis (4, 16, 36, 50; second section ofRESULTS). Both of these synthetic activities wouldrequire active energy metabolism in the Hfrparent at the inception of mating (first section ofRESULTS). This DNA synthesis in the Hfr parentwould act as the driving force for the initiation ofchromosome transfer, with the introduction ofseveral per cent of the leading extremity of thedonor chromosome into the recipient parent(second and fourth sections of RESULTS). Moreextensive chromosome transfer would then re-quire the establishment of effective homologouspairing between the lead region of the donorchromosome and the comparable segment of therecipient chromosome (fourth section of RE-SULTS). The F- parent would then wind in thedonor chromosome with an expenditure of energy(20; first section of RESULTS). This process wouldensure that homologous regions of the donor andrecipient chromosomes would be brought intosynaptic union to provide for the known orderlyand efficient integration of donor genetic materialinto recombinants.The above-described events and parental ac-

tivities during bacterial conjugation are supportedby many experimental observations, and we arenot aware of any findings which would appreci-ably alter the model as proposed. Certain otheraspects of bacterial conjugation, especially thoseconcerned with the mechanism of chromosometransfer, are not yet resolved, since contradictingresults and interpretations of data do exist. Thedependency of chromosome transfer on continualDNA synthesis in the Hfr parent as suggested byJacob, Brenner, and Cuzin (38) is not supportedby our findings or by those of Pritchard (48).On the other hand, there is evidence which hasbeen interpreted to indicate that replication of thedonor chromosome normally accompanies trans-fer and that this replication occurs in the Hfr

VOL. 32, 1968 345

on Septem


mbr.asm

.org/D

ownloaded from


CURTISS ET AL.

parent (6, 33, 34, 49). Philosophically, we mightwish that what normally occurs would indicatewhat must of necessity occur. However, we believethat chromosome transfer is usually a cooperativeeffort by both mating partners and that alterna-tive means for accomplishing the same end resultmight have been developed during the course ofbacterial evolution. Our finding that the F-parent is responsible, at least in part, for con-trolling the rate of chromosome transfer wouldseem to eliminate the requirement that the donor-mediated replication of its chromosome duringtransfer controlled the rate of transfer as firstsuggested by Jacob, Brenner, and Cuzin (38).Certainly, replication of the DNA strand retainedin the Hfr cell during chromosome transfer wouldensure the genetic viability of the donor cell, sincein the absence of DNA synthesis mating wouldoften be a suicidal act for the Hfr parent.There also exist in the literature and in this

paper conflicting data and contradictory conclu-sions or- metabolic activities occurring in the F-parent during chromosome transfer. We havedemonstrated an association between energy me-tabolism in the F- parent and control of the rateof chromosome transfer and a requirement foreffective homologous pairing between the leadregion of the Hfr chromosome and the com-parable segment of the recipient chromosome toachieve a substantial amount of chromosometransfer. It would therefore appear that the mostplausible explanation for the results obtained withamino acid-starved F- strains would be thatcompletion of vegetative chromosome replicationmust occur in the F- prior to the winding in ofthe donor chromosome. A corollary suggestionwould be that chromosome transfer and vegeta-tive replication of the recipient chromosomemight be mutually exclusive processes. A predic-tion of the above hypothesis is that prevention ofDNA synthesis in a logarithmically growing F-population should markedly reduce the numberof F- cells capable of involving themselves inchromosome transfer. While Bonhoeffer's (7)results are in complete accord with this predic-tion, the results of thymine starvation of thy- F-strains are not (48; second section of REsuLTS).Further experimentation is now in progress in thehope of resolving these problems and to test thehypothesis that completion of vegetative chromo-some replication in the F- is necessary forchromosome transfer.

[Note added in proof. During this past summer,we (Curtiss and Williams, unpublished data) con-ducted a series of experiments on the effects ofamino acid starvation on the F- parent. Theresults obtained strongly suggest that chromo-

some transfer is not dependent upon completionof an old round or initiation of a new round ofvegetative replication of the F- chromosome.The cause of the effects described in the thirdsection of the Results is as yet unknown.]Another problem associated with metabolic

activities occurring in the F- parent duringconjugation concerns the effects of purine starva-tion of pur- F- strains. Freifelder (30) demon-strated that purine starvation depressed the fre-quency of episome and chromosome transfer,whereas Gross and Caro (34) observed no effectof purine starvation or UV irradiation of the F-parent on chromosome transfer. We (Curtiss andMays, unpublished data) have partially confirmedFreifelder's results, but have also noted thatpurine starvation inhibits certain events involvedin the integration of transferred donor geneticmaterial. We also suggest the possibility thatcertain treatments of the F- parent, such as purinestarvation, might involve restriction of normallytransferred donor genetic material. Gross andCaro (34) used E. coli C as their F- recipient,and this strain is normally incapable of restriction.The conclusions drawn and the suggestions

made on the basis of the data presented in thismanuscript have been substantiated in part bythe studies of Cohen et al. (15, 16) on the physicalproperties of DNA transferred by F+, F', andHfr donors to the DNA-less minicells isolatedand described by Adler et al. (3). This researchhas also permitted a more precise definition ofcertain steps and features of the conjugal transferof genetic information from donor to recipientstrains of E. coli K-12.

ACKNOWLEDGMENTSWe express our thanks to David Allison for per-

forming the electron microscopy and F pili countsand to our colleagues H. I. Adler, L. G. Caro, andA. Cohen for critical reading of this manuscript andfor many helpful suggestions during the course ofthis work.

This research was sponsored by the U.S. AtomicEnergy Commission under contract with Union Car-bide Corp. Research conducted by the senior authorat the University of Chicago was done during thetenure of a Public Health Service Predoctoral Fellow-ship and was supported by Public Health ServiceResearch Training Grant 5T1 GM-603.

LITERATURE CrTED1. Adelberg, E. A., and S. N. Burns. 1960. Genetic

variation in the sex factor of Escherichia coli.J. Bacteriol. 79:321-330.

2. Adelberg, E. A., and J. Pittard. 1965. Chromo-some transfer in bacterial conjugation. Bac-teriol. Rev. 29:161-172.

3. Adler, H. I., W. D. Fisher, A. Cohen, and A. A.

346 BACTERIOL. REV.

on Septem


mbr.asm

.org/D

ownloaded from



Hardigree. 1967. Miniature Escherichia colicells deficient in DNA. Proc. Natl. Acad. Sci.U.S. 57:321-326.

4. Barbour, S. D. 1967. Effect of nalidixic acid onconjugational transfer and expression of epi-somal lac genes in Escherichia coli K12. J.Mol. Biol. 28:373-376.

5. Berg, C. M., and R. Curtiss. 1967. Transpositionderivatives of an Hfr strain of Escherichia coliK-12. Genetics 56:503-525.

6. Blinkova, A. A., S. E. Bresler, and V. A. Lanzov.1965. DNA synthesis and chromosome trans-fer in Escherichia coli K-12. Z. Vererbungslehre96:267-274.

7. Bonhoeffer, F. 1966. DNA transfer and DNAsynthesis during bacterial conjugation Z.Vererbungslehre 98:141-149.

8. Bouck, N., and E. A. Adelberg. 1963. The rela-tionship between DNA synthesis and conjuga-tion in Escherichia coli. Biochem. Biophys.Res. Commun. 11:24-27.

9. Brinton, C. C., Jr. 1965. The structure, function,synthesis and genetic control of bacterial piliand a molecular model for DNA and RNAtransport in gram-negative bacteria. Trans.N.Y. Acad. Sci. 27:1003-1054.

10. Brinton, C. C., Jr. 1967. Contributions of pih tothe specificity of the bacterial surface, and aunitary hypothesis of conjugal infectiousheredity, p. 37-70. In B. D. Davis and H. J.Vogel (ed.). The specificity of cell surfaces.Prentice-Hall, Englewood Cliffs, N.J.

11. Brinton, C. C., Jr., P. Gemski, Jr., and J. Carna-han. 1964. A new type of bacterial pilus geneti-cally controlled by the fertility factor of E. coliK12 and its role in chromosome transfer.Proc. Natl. Acad. Sci. U.S. 52:776-783.

12. Broda, P. 1967. The formation of Hfr strains inEscherichia coli K-12. Genet. Res. 9:35-47.

13. Cairns, J. 1963. The chromosome of Escherichiacoli. Cold Spring Harbor Symp. Quant. Biol.28:43-46.

14. Cohen, A., D. P. Allison, H. I. Adler, and R.Curtiss. 1967. Genetic transfer to mini cells ofEscherichia coil K-12. Genetics 56:550-551.

15. Cohen, A., W. D. Fisher, R. Curtiss, and H. I.Adler. 1968. DNA isolated from Escherichiacoil minicells after mating with F+ cells. Proc.Natl. Acad. Sci. U.S. 61:61-68.

16. Cohen, A., W. D. Fisher, R. Curtiss, and H. I.Adler. 1968. The properties of DNA trans-ferred to minicells by conjugation. ColdSpring Harb. Symp. Quant. Biol., in press.

17. Curtiss, R. 1964. A stable partial diploid strainof Eseherichia coli. Genetics 50:679-694.

18. Curtiss, R. 1965. Chromosomal aberrationsassociated with mutations to bacteriophageresistance in Escherichia coli. J. Bacteriol.89:28-40.

19. Curtiss, R. 1968. Ultraviolet-induced geneticrecombination in a partially diploid strain ofEscherichia coli. Genetics 58:9-54.

20. Curtiss, R., and L. J. Charamella. 1966. Role of

the F- parent during bacterial conjugation inEscherichia coli. Genetics 54:329-330.

21. Curtiss, R., L. J. Charamella, C. M. Berg, andP. E. Harris. 1965. Kinetic and genetic analysesof D-cycloserine inhibition and resistance inEscherichia coli. J. Bacteriol. 90:1238-1250.

22. Curtiss, R., and D. R. Stallions. 1967. Energyrequirements for specific pair fonnation duringconjugation in Escherichia coli K-12. J. Bac-teriol. 94:490-492.

23. Curtiss, R., D. R. Stallions, J. A. Mays, and J.Renshaw. 1967. Mechanism of chromosomemobilization and transfer by F+ donors ofEscherichia coil K-12. Genetics 56:553-554.

24. Cuzin, F., and F. Jacob. 1966. Inhibition par lesacridines du transfert genetique par les souchesdonatrices d'Escherichia coli K12. Ann. Inst.Pasteur 111:427-436.

25. Cuzin, F., and F. Jacob. 1967. Mutations del'episome F d'Escherichia coli K12. HI. Mu-tants a replication thermosensible. Ann. Inst.Pasteur 112:397-418.

26 De Haan, P. G., and J. D. Gross. 1962. Transferdelay and chromosomal withdrawal duringconjugation in E. coli. Genet. Res. 3:251-272.

27. Demerec, M., E. A. Adelberg, A. J. Clark, andP. E. Hartman. 1966. A proposal for a uni-form nomenclature in bacterial genetics.Genetics 54:61-76.

28. Fisher, K. W. 1957. The nature of the endergonicprocesses in conjugation in Escherichia coilK-12. J. Gen. Microbiol. 16:136-145.

29. Fisher, K. W. 1966. Amino acid deprivation andits effect on mating ability in Escherichia coilK-12. Genet. Res. 8:115-118.

30. Freifelder. D. 1967. Role for the female in bac-terial conjugation in Escherichia coli. J. Bac-teriol 94:396-402.

31. Freifelder, D. R., and D. Freifelder. 1968. Stud-ies on Escherichia coli sex factors. H. Somephysical properties of F' lac and F DNA. J.Mol. Biol. 32:25-35.

32. Glansdorff, N. 1967. Pseudoinversions in thechromosome of Escherichia coil K-12. Ge-netics 55:49-61.

33. Gross, J. D., and L. G. Caro. 1965. Genetictransfer in bacterial mating. Science 150:1679-1684.

34. Gross, J. D., and L. G. Caro. 1966. DNA trans-fer in bacterial conjugation. J. Mol. Biol.16:269-284.

35. Hayes, W. 1964. The genetics of bacteria andtheir viruses, studies in basic genetics andmolecular biology. Blackwell Scientific Publi-cations, Oxford.

36. Hollom, S., and R. H. Pritchard. 1965. Effect ofinhibition of DNA synthesis on mating inEscherichia coli K-12. Genet. Res. 6:479-483.

37. Jacob, F., and E. A. Adelberg. 1959. Transfert decharacteres g6n6tique par incorporation aufacteur sexual d'Escherichia coil. Compt. Rend.249:189-191.

38. Jacob, F., S. Brenner, and F. Cuzin. 1963. On

347VOL. 32, 1968

on Septem


mbr.asm

.org/D

ownloaded from


348 CURTIS

the regulation of DNA replication in bacteria.Cold Spring Harbor Symp. Quant. Biol. 28:329-348.

39. Jacob, F., and E. L. Wollman. 1961. Sexualityand the genetics of bacteria. Academic Press,Inc., New York.

40. Johnson, E. M., S. Falkow, and L. S. Baron.1964. Recipient ability of Salmonella typhosain genetic crosses with Escherichia coli. J. Bac-teriol. 87:54-60.

41. Krisch, R. E., and M. J. Kvetkas. 1966. Inhibi-tion of bacterial mating by amino acid depriva-tion. Biochem. Biophys. Res. Commun. 22707-711.

42. Lark, K. G., T. Repko, and E. Hoffman. 1963.The effect of amino acid deprivation on sub-sequent deoxyribonucleic acid replication.Biochim. Biophys. Acta 76:9-24.

43. Lederberg, E. M., and J. Lederberg. 1953. Ge-netic studies of lysogenicity in Escherichia coli.Genetics 38:51-64.

44. Lennox, E. S. 1955. Transduction of linked ge-netic characters of the host by bacteriophageP1. Virology 1:190-206.

45. Low, B. 1966. Low recombination frequency formarkers very near the origin in conjugation inE. coli. Genet. Res. 6:469-473.

46. Maal0e, O., and P. C. Hanawalt. 1961. Thyminedeficiency and the normnal DNA replicationcycle. J. Mol. Biol. 3:144-155.

47. Pittard, J., and E. M. Walker. 1967. Conjugationin Escherichia coli: recombination events interminal regions of transferred deoxyribo-nucleic acid. J. Bacteriol. 94:1656-1663.

48. Pritchard, R. H. 1965. The relationship betweenconjugation, recombination, and DNA syn-thesis in Escherichia coil, p. 55-78. In Geneticstoday, Proc. Intern. Congr. Genet., 11th, TheHague, 1963. Pergamon Press, London.

S ET AL. BACTERIOL. REv.

49. Ptashne, M. 1965. Replication and host modifi-cation of DNA transferred during bacterialmating. J. Mol. Biol. 11:829-838.

50. Roeser, J., and W. A. Konetzka. 1964. Chromo-some transfer and the DNA replication cyclein Escherichia coli. Biochem. Biophys. Res.Commun. 16:326-331.

51. Scaife, J., and A. P. Pekhov. 1964. Deletion ofchromosomal markers in association withF-prime factor formation in Escherichia coliK-12. Genet. Res. 5:495-498.

52. Schwartz, N. M. 1963. Nature of ethyl methane-sulfonate induced reversions of lac- mutantsof Escherichia coli. Genetics 48:1357-1375.

53. Silver, S. D. 1963. The transfer of material dur-ing mating in Escherichia coli. Transfer ofDNAand the upper limnits on the transfer of RNAand protein. J. Mol. Biol. 6:349-360.

54. Suit, J. C., T. S. Matney, C. 0. Doudney, andD. Billen. 1964. Transfer of the Escherichiacoli K-12 chromosome in the absence of DNAsynthesis. Biochem. Biophys. Res. Commun.17:237-241.

55. Taylor, A. L., and C. D. Trotter. 1967. Revisedlinkage map of Escherichia coli. Bacteriol. Rev.31:332-353.

56. Tomizawa, J., and N. Anraku. 1964. Molecularmechanisms of genetic recombination in bac-teriophage. I. Effect of KCN on genetic re-combination of phage T4. J. Mol. Biol. 8:508-515.

57. Tomizawa, J., and N. Anraku. 1964. Molecularmechanisms of genetic recombination in bac-teriophage. II. Joining of parential DNA mole-cules of phage T4. J. Mol. Biol. 8:516-540.

58. Wu, T. T. 1967. Recombination frequencies ofproximal markers in bacterial conjugation. J.Theoret. Biol. 17:40-46.

on Septem


mbr.asm

.org/D

ownloaded from


Documents

Dec. Vol. In Parental Functions During Conjugation Escherichia … · macromolecular syntheses, or both, in one or both parents. These experiments were done to determine which parent