the generation of concatemeric plasmid dna in bacillus subtilis as

© 7990 Oxford University Press Nucleic Acids Research, Vol. 18, No. 16 4651

The generation of concatemeric plasmid DNA in Bacillussubtilis as a consequence of bacteriophage SPP1 infection

Alicia Bravo and Juan C. Alonso*Max-Planck-lnstitut fur Molekulare Genetik, Ihnestrasse 73, D-1000 Berlin 33, FRG

Received July 3, 1990; Accepted July 23, 1990

ABSTRACT

Bacteriophage SPP1 infection of Bacillus subtilis cellsbearing plasmids induces the synthesis ofmultigenome-length plasmid molecules. Twoindependent pathways can account for this synthesis.In one of those, homology to the phage genome isrequired, whereas in the other such homology is nota prerequisite. In wild type cells both modes overlap.In dnaB(Ts), at non permissive temperature, or in recEpolk strains the main concatemeric plasmid replicationmode is the homology-dependent £lasmid (hdp) mode.The rate of recombination-dependent concatemericplasmid DNA synthesis is a consequence of a phage-plasmid interaction which leads to chimericphage::plasmid DNA. The second mode, which is an/romology-independent plasmid (hip) mode seems tobe triggered upon the synthesis of a phage encodedproduct(s) (e.g. inactivation of the exonuclease Venzyme).

INTRODUCTION

Plasmid DNA, which is usually smaller in size than the DNAof a bacteriophage genome, can be encapsidated into phageproheads to produce plasmid transducing particles (1). Theencapsidation of linear head-to-tail plasmid molecules has beendetected in all bacteria tested so far (2-9) . The frequency ofplasmid containing phage particles is greatly enhanced [102 tolO^fold] when DNA homology between plasmid and phage isprovided (6, 7). Beyond this observation, however, the molecularevents that trigger the synthesis of packagable concatemericplasmid DNA are not well understood. Depending on thereplication machinery eventually used, two models to explain sucha phenomenon were proposed. In model 1, a phage encodedproduct is inhibitory for termination of plasmid replication,leading to the formation of plasmid multimers (7). In this case,homology has the effect of introducing, via homologousrecombination between phage and plasmid concatemericmolecules, the phage packaging recognition signal (pac site) intothe plasmid concatemer. The model implies that synthesis of theplasmid concatemer is homology-independent, but depends onplasmid-initiated replication. In model 2, a single-stranded endof phage DNA invades a homologous region of a double-stranded

plasmid molecule. This recombination intermediate is thenconverted into a replication fork (10, 11). Such a synapsisbetween a phage replication intermediate and a plasmid moleculewould lead to the generation of a phage: .plasmid DNA chimeracarrying the pac site. In this situation the phage replisome isdriving the synthesis of concatemeric plasmid DNA (6, 9).

In certain Bacillus subtilis and Escherichia coli strains, linearmultigenome-length plasmid molecules,as those described above,were observed 'independent of phage infection'. An impairmentin the exonuclease V (ExoV) activity [also termed RecBCD inE. coli or Add ABC in B. subtilis (12)] triggers the synthesis ofconcatemeric plasmid molecules (13-16). This synthesis requiresthe active RecA/RecE (equivalent proteins from E. coli and B.subtilis origin, respectively) and DNA polymerase I products (13,14, 16). This novel form of plasmid DNA synthesis was reportedto be dependent on functions involved in DNA recombinationbut independent of plasmid replication initiation functions orDnaA protein as in the case of oriC plasmids (14, 15). Consistentwith this is the fact that such concatemeric plasmid DNA synthesisis partially resistant to chloramphenicol and rifampicin (17).Furthermore, for phage concatemeric DNA synthesis,inactivation of the ExoV enzyme is needed (18, 19, 20).

In this communication, we analyze the replication of plasmidswith or without a DNA fragment from SPP1 genome in SPP1infected B. subtilis cells. Our results suggest the existence of twooverlapping concatemeric plasmid replication modes followingphage infection. Such plasmid replication will be referred to ashomology-independent or homology-dependent when thepresence of the SPP1 DNA fragment, on the plasmid, has aninfluence on the rate of concatemeric plasmid replication.Furthermore, we show that phage::plasmid chimeras, carryingthe pac recognition signal, can be generated during phageinfection.

MATERIALS AND METHODS

Bacterial strains, phages and plasmids

B. subtilis strains YB886 (wild-type) and the isogenic YB1015(recEA), BG219 (recE6 polA5), BG193 (dnaB37) and BG195(dnaBl9) have been previously reported (21, 22). The phagesused were SPP1 wild type and SPPlaml09 (23).

The B. subtilis plasmids used were based either on the pC194

* To whom correspondence should be addressed

Downloaded from https://academic.oup.com/nar/article-abstract/18/16/4651/1080552by gueston 13 April 2018

4652 Nucleic Acids Research, Vol. 18, No. 16

or pUBl 10 replicons. Plasmids pC1941, pl958 and pBG59 werepreviously described (24, 6). Plasmids pi934 and pi936 arepC1941 derivatives in which a pac site containing SPP1 DNAfragment of 840 bp or 197 bp, respectively, was introduced.pBD331 contains a 2.1 kb DNA fragment adjacent to the SPP1pac region, but lacks the pac site (24).

Transduction and transformationTransducing phage stocks were prepared and transduction wasperformed as previously described (24). Transformation followedthe protocol of Rottlander and Trautner (25). Antibiotic resistanttransductants or transformants were selected on TY platescontaining 5 /tg/ml of chloramphenicol or 35 /tg/ml of ampicillin.

Analysis of plasmid DNA formsB. subtilis strains YB886 and BG219 bearing plasmids weregrown at 37°C to an OD560 of 0.45 (about 5.0xlO7 cells/ml)in TY medium supplemented with the appropriate antibiotic, andinfected with 5 phage/bacterium. Phage addition marked the timezero of our experiments. At given times, aliquots of equal volumewere removed and poured onto frozen TBT (100 mM Tris-HClpH 7.5, 100 mM NaCl, 10 mM MgCl) supplemented withsodium azide to 10 mM. In temperature shift experiments, thestrain BG193 dnaB(Ts) bearing plasmids was grown at 30°C,to an OD560 of 0.45 and transferred to 46°C. 10 min after shiftup the cells were infected with 5 phage/bacterium and processedas described above.

Crude cell lysates were obtained and processed as previouslydescribed (14). Since the events to be monitored account for lessthan 1 % of total DNA, prior to electrophoresis to avoid DNAsaturation in the nylon membrane, to gain sharpness in thetransferred bands and linearity in our quantifications, the DNApreparations were treated with a restriction enzyme which digestphage and chromosomal DNA, but not plasmid DNA. Aliquotsof crude cell lysates were electrophoresed overnight in 0.8%agarose. The DNA was transferred to a nylon membrane(GeneScreen, NEN) and processed essentially as described byManiatis et al. (26).

The relative amounts of plasmid DNA forms given in the textrefer to the intensities of the plasmid bands present on theautoradiograms. They were quantitatively scanned with a laserdensitometer (LKB UltroScan XL). The linearity of the responsewith respect to DNA concentration was checked usingautoradiograms of different exposure times. Quantitative scanswere integrated using the LKB GelScan XL software package.

Since plasmids pC1941 and pl958 shared the pC194 replicon(equal copy number ?) and the rate of SPP1 DNA synthesis(correction for the amount of infected cells) in cells carrying eitherplasmid pC1941 or pl958 does not differ more than 1.4-fold weassumed that normalizing the amount of concatemeric plasmidDNA form to the amount of form I (monomeric and dimericsupercoiled covalently closed plasmid) DNA could represent thein vivo situation.

Other biochemical techniquesPlasmid DNA was prepared on preparative and analytical scalesas previously described (27). Restriction enzymes and otherrelevant enzymes were purchased from Boehringer (Mannheim,FRG). All enzymes were used as described by the suppliers.Other standard techniques as DNA hybridization and nick-translation of DNA with [«32P]dATP [Amersham-Buchler(UK)] were performed as described by Maniatis et al. (26).

Radioactive probes were prepared by nick-translation of the E.coli plasmids pBT39 (caf-probe, containing 0.7 kb of pC1941),pBT38 (We-probe, containing 0.5 kb of pUBHO) and pBT43(pac-probe, containing 0.7 kb of the SPP1 pac region). Theprobes detect the chloramphenicol (cat) and phleomycin (ble)resistance genes of plasmids pC 1941 and pUBl 10 and the originof SPP1 packaging (pac), respectively.

• SPP1 - S P P 1 B SPP1 -SPP1

- . J

0 5 1015 2025 30 0 10 20 3050mm 0 5 1 0 1 5 2 0 2 5 3 0 0 10 20 30 50 min

Fig. 1. Time course of the relative plasmid DNA mass accumulated during SPP1infection on wild type cells carrying plasmid pC1941 (A) or pl958 (B). Culturesof YB886 cells harboring plasmid were grown at 37 °C to OD560 = 0.45 andhalf of the culture was infected with SPP1 (time zero). From the infected ( +SPP1) and non-infected (-SPP1) cultures samples were withdrawn at indicatedtimes. The vertical arrows denote the DNA preparations, obtained 30 min afterinfection, digested either with EcoRl (A) or Thai (B). The horizontal arrowsindicate the position of form I plasmid DNA and the origin of the gel.

Fig. 2. Kinetics of accumulation of concatemeric plasmid DNA during SPP1infection of wild type (A) and dnaBtf (B) strains. The relative amount ofconcatemeric plasmid DNA is presented as a ratio of concatemeric to form I(monomeric and dimeric) plasmid DNA. The intensities of plasmid bands presenton the autoradiograms, after quantitative scann with a laser densitometer, wereintegrated by the use of the GelScan software package. ( • ) plasmid pC 1941 and(LI) plasmid pl958.


Nucleic Acids Research, Vol. 18, No. 16 4653

RESULTSSynthesis of multigenome-length plasmid molecules duringSPP1 infectionIn order to understand how plasmid transducing particles aregenerated and which role DNA homology plays in this process,we began with the characterization of plasmid DNA forms duringphage infection. For this analysis, two plasmids presenting eitherno apparent (pC1941) or extensive (174 base pairs, pl958) DNAhomology to the SPP1 genome were compared. pC1941 istransduced with a frequency of about 10~5 trans-ductants/surviving cell, whereas pl958 is transduced with afrequency of about 10~3 transductants/surviving cell (6).

B. subtilis YB886 cells carrying plasmids pC1941 or pl958were grown exponentially at 37°C to about 5.0X107 cells/mland either infected with SPP1 or allowed to grow further (non-infected cells). At different times, samples from infected and non-infected cultures were collected and total intracellular DNA waspurified. The DNA preparations were treated with HpaYL orHindHL, in the case of pC1941 or pl958, respectively. Theseenzymes have recognition sites in the phage and hostchromosome, but not in plasmid DNA. The DNA forms werethen separated electrophoretically and analyzed by Southernhybridization using the cat plasmid probe.

As shown in Figure 1, after SPP1 infection a slowly movingDNA form, which hybridizes specifically to the car-probe, isaccumulating (Fig. 1A and IB). This plasmid DNA form co-migrates with undigested phage DNA (data not shown). Whenthe total intracellular DNA is also digested with an enzyme thatcleaves the plasmid once (£coRI and Thai for pC1941 and pl958DNAs, respectively) only one signal corresponding to linear unit-length plasmid molecules is observed (Fig. 1A and IB, lane 8).

B

Thus, the slow migrating DNA form is of plasmid origin andrepresents a plasmid concatemer which must be organized in ahead-to-tail configuration.

Such plasmid concatemeric DNA was neither detected inuninfected cells (Fig. 1A and IB) nor following infection witha pleiotropic suppressor-sensitive SPP1 mutant (amlO9) (data notshown), which synthesizes only a few early phage products undernon permissive conditions (see 23). Thus, this de novo plasmidconcatemeric DNA synthesis is dependent on phage encodedproduct(s).

The amount of plasmid concatemer detected 30 min afterinfection is about 15- to 17-fold greater than the amount detected15 min after infection (Fig. 2A). The synthesis of concatemericplasmid form has the same kinetics whether the plasmid shareshomology or not with the phage DNA (Fig. 2A). Since phageconcatemeric DNA synthesis implies inactivation of the ExoVenzyme (18, 19, 20) and, in both E. coli and B. subtilis, linearhead-to-tail plasmid molecules have been detected after ExoVinactivation (13, 14), we interpret that SPP1 promotes thesynthesis of such concatemeric plasmid molecules by inactivatingthe ExoV enzyme.

In spite of the same amount of concatemeric plasmid DNA

Table I: Influence of the DnaB product on plasmid transduction frequencies atnonpermissive temperature

Donor8

strain

YB886

BG193

BG195

Relevantgenotype

dna +

dna&n

dnd&\9

Time of infectionafter shift (min)

0300

300

30

Transductants/surviving cells

6.6X10"3

5.1xlO~3

7.0xl0~ 3

4.5X10"3

7.2X10"3

4.7X10"3

(a) The strains containing plasmid pl958 were grown at 30°C. After such growthone half was infected with SPP1 and kept at 46°C till lysis and the other onewas incubated for a further 30 min at 46°C prior to infection. The lysates wereused to transduce YB 886 cells at 37 °C.

B

I•

0 5 10 IS 20 25 30 mm 05 10 15 20 25 30 mm

Fig. 3. Effect of the dnd&yi mutation on the accumulation of concatemeric plasmidmolecules. Cultures of BG193 carrying plasmid pC1941 (A) or pl958 (B) weregrown at 30°C to OD560 = 0.45 and transferred to 46°C. 10 min aftertemperature shift up, the cultures were infected with SPP1 (time zero).Experimental conditions are like in Fig. 1.

0 5 X) 15 20 25 30mm 0 5 101520 25 30mm

Fig. 4. Effect of the recE6polA5 mutations on the accumulation of concatemericplasmid molecules. Cultures of BG219 carrying plasmid pUBl 10 (A) or pBG59(B) were grown at 37°C to OD560 = 0.45 and infected with SPP1 (time zero).Experimental conditions were those of Fig. 1.



B§ £ £a. a. o. H E E l

j i i T

7 .1-

4 .8 -

3.4-

2.7 -

1.9-

1.2-

0.7-

#tA

i

E1

5kbA

1

EI

A

|_

1.42

3.87kb

A

H

U 8

1.06

A

II

1.06

E1

2,1

3.9 kb

1.97

2.A1

» 1 «

II

EI

69 kb

1.97

I

\

I

\

l° '2|I

pBO331

p1934

p 1936

3.23kb 2k b

Fig. 5. Hybridization analysis of encapsidated plasmid DNA. (A) DNA from SPP1 particles grown on wild type cells bearing the indicated plasmids was digestedwith Aflll and analyzed by Southern using nick-translated pC1941 DNA. (B) Physical map of the phage pac region and concatemeric plasmid molecules. Open andclosed bars represent phage and plasmid DNAs, respectively. The arrow and the closed square indicate the phage pac site. Abbreviations: A, AflU; E, £coRI; H, HindSl.

synthesized in the absence or presence of homology, the strikingdifference between the transduction frequencies in the two casesremain. pl958 (6.4xlO~3 transductants/ surviving cell) istransduced with a 230-fold higher efficiency than pC1941(2.7xlO~5 transductants/surviving cell) (6).

Homology-dependent plasmid replication during SPP1infectionIt is known that phage SPP1 and pC1941 or pi958 plasmid DNAreplication are independent of the B. subtilis DnaB protein whichis a component of the prereplisome (23, 16). However, the hostDnaB protein is required for the initiation of chromosomal DNAreplication (28) as well as for multigenome-length plasmid DNAreplication (17, Viret and Stiege, personal communication).Furthermore, in the latter DNA synthesis the recE (equivalentto recA of E. coll) and polA gene products are involved (13-16).In order to test whether the synthesis of multigenome-lengthplasmid molecules following SPP1 infection is only due to anindirect effect [replication as a consequence of inactivation ofthe ExoV enzyme (13, 14)] or due to some other mode of plasmidreplication (see Introduction, model 2), the rate of concatemericplasmid DNA synthesis following SPP1 infection was measuredin different genetic backgrounds.

B. subtilis dnaB31(Ts) cells bearing plasmid pC1941 or pl958were shifted from 30°C to 46°C. 10 min after the shift to nonpermissive temperature, cultures were infected with bacteriophageSPP1. At given times samples were collected, total intracellularDNA was purified and processed as described above.

After the temperature shift, chromosomal replication wasinhibited as expected for a dnaB mutant strain (data not shown).The amount of slowly moving pC1941 DNA detected in thednaB(Ts) strain (Fig. 3A) is markedly reduced and followsdifferent kinetics (Fig. 2) when compared to the amount detected

in the wild type strain (Fig. 1A). In contrast, slowly movingpl958 DNA molecules are detected about 10 min after infection(Fig. 3B) and its kinetics of synthesis is similar in dnaB and wildtype cells (Fig. 2). Thus, a) thermal inactivation of the DnaBproduct affects /zomology-mdependent concatemeric /?lasmidDNA synthesis (hip), and ii) homology between the phage andthe plasmid DNA compensate for the absence of the DnaBproduct. Hence, we termed this mechanism /lomology-dependentplasmid Quip) replication mode.

The amount of concatemeric plasmid DNA 20 min afterinfection is about 28-fold higher for pi958, than in the case ofpC1941 (Fig 2B). However, at later times pC1941 concatemericDNA accumulates (Fig. 2B). This limited accumulation ofpC1941 concatemeric DNA might be due to leakiness of thednaB37(Ts) mutation. Alternatively, the slower-moving plasmidDNA form in cells bearing plasmid pC1941 could be synthesizedat a lower efficiency by a hdp mechanism due to a short stretchof natural homology between the phage and the plasmid (6). 30min after infection the amount of head-to-tail concatemericplasmid DNA is still more than 8-fold higher in pl958 than inpC1941.

pi958 transduction frequency of phage stock lysates amplifiedat 46°C in the wild type, dnaB37(Ts) or dnaB\9(Js) strainsbearing plasmid does not differ more than two-fold (see Table I).

To confirm whether homology plays any role in the synthesisof concatemeric plasmid DNA, we have used the recE6 polA5strain. Since, the recE6 polAS strain by construction is resistantto chloramphenicol (recE6 = recE::cat insertional inactivation)(22), the neomycin-resistant plasmids pUBHO (no homology)and pBG59 (with homology) were used. Transduction in a wildtype background of such plasmid DNA by phage SPP1 is identicalto the one observed with the pC 1941 derived replicons (6). Here,BG219 (recE6 polA5) cells bearing plasmid pUBl 10 or pBG59



MWM

IMWM

- 6

- 3 . 5

,-1.9

-1.2

- 0 . 7 3

-1.9

-1.2

- 0 . 7 3

1 2

mm1 2

A

A

K

t1

1.10 -

E1

Pv

Pv

—+-

H He-U

He

:

A

-1.05 - t~ -1.03-

H A

He

PvI

: He|

+—0.72 —

Pv

i

1 0.96 4 0.93 -

Fig. 6. Southern analysis of SPP1 infected B. subtilis cells carrying plasmid pl958.(A) Total DNA obtained 30 min after infection of YB886 cells was digested withEcoRl and Aflll (1) or EcoRl and Pvull (2), and blot hybridized with the catprobe. The arrows indicate the position of minor bands (1.9 and 3.0 kb) andthe asterisk indicates a band of 1.7 kb. (B) Total DNA obtained 20 min (1,3and 4) and 30 min (2, 5 and 6) after infection of strains YB886 (1, 2, 3, and5) or BG193 (4 and 6) were digested with Aflll (1 and 2) or Hindlll (3, 4, 5and 6) and analyzed using thepac probe. Arrows denote the minor band of 3.8kb and the position of the HindUl resistant DNA molecules. The molecular weightmarker (MWM) is SPP1 EcoRl digested. (C) Structure of phage: :plasmid chimerasdetected in SPP1 infected cells. Open bars represent phage DNA, closed barsdenote concatemeric plasmid DNA, the striped box shows the homologous regionbetween the phage and the plasmid. The arrow and the closed square indicatethe phage pac site. The wavy lines under the phage and plasmid molecules denotetheir respective probes (pac and cat probes). The lengths of the DNA segments(in kb) are indicated. Abbreviations: A, Aflll; E, EcoRl; H, //indffl, He, HindU;Pv, Pvull.

were infected with SPP1. At different times total DNA waspurified, separated on agarose gel, blotted and hybridized withthe ble plasmid probe.

The appearance of pUBl 10 concatemeric DNA molecules isdelayed and its amount markedly reduced when compared to theamount synthesized in the presence of homology (Fig. 4). Theamount of the pBG59 concatemeric DNA detected 20 min afterinfection is about 7-fold higher than when homology is notpresent. The kinetic of synthesis of pBG59 concatemeric DNAis similar to the one observed for pi958 in wild type and dnaBiJs)

strains (see Figs. 1-3). Thus, independent of the inactivationof the recE and polA functions. Furthermore, the pBG59transduction frequency obtained with the stock lysates preparedin the recE6 polA5 strain does not differ more than two-fold whencompared with the rec+ cells (YB886) (data not shown).

From those data we can conclude that the synthesis ofconcatemeric phage, hdp and hip DNA replication are closelylinked in time, but hip replication could become unlinked eitherby thermal inactivation of the DnaB product or by mutations inthe recE and polA genes. DNA homology between the plasmidand the phage is necessary for hdp replication.

Phage:: plasmid chimeras are generated during the synthesisof concatemeric plasmid DNAAll virus packaging apparatus recognize the nucleic acid to bepackaged by virtue of a specific base sequence in such nucleicacid [packaging recognition signal (pac site) (29)]. In SPP1 DNAhomology between the plasmid and phage genomes, irrespectiveof whether this includes the pac signal on the plasmid, greatlyincreases the amount of concatemeric plasmid DNA encapsidatedinto phage proheads (24). How does the phage packagingmachinery recognize the plasmid concatemer as a substrate forencapsidation? A simple assumption is that, in a 'homology-dependent fashion', a concatemeric plasmid::phage chimericmolecule is generated. Such hybrid molecule would contain thephage pac site. In the case of SPP1, packaging of such chimerawould proceed sequentially and unidirectionally from the pac site(35, 31). To test the above possibility, stock lysates of wild typecells bearing pC 1941-derived plasmids containing SPP1 DNAfragments from regions near pac [pac-free (pBD331) and pac-containing (pl934 and pl936)plasmids] were obtained. Totalpackaged DNA was purified, digested with AflO. (cleaves theplasmids once, but does not cleave phage DNA), and analyzedby Southern blot using as probe nick-translated pC1941 DNA.In addition to a major DNA band, which corresponds to unit-length plasmid molecules, a discrete minor band was detected(Fig. 5A). The length of the minor band is 3.9 kb, 2.7 kb or2.0 kb for plasmids pBD331, pl934 or pl936, respectively. Thesize of such fragments suggest that in all cases the phagepackaging signal is utilized (see Fig. 5B). Since pBD331 doesnot have a pac site, we infer that the plasmid gained the pac sitefrom the phage. A similar observation was reported for theStaphylococcus aureus phage <£ll-pT181 system (7). Theseresults reflect that, even in the absence of a pac site in the regionof plasmid/phage homology, concatemeric plasmid::phagechimeras carrying pac must be generated during SPP1 infection.

Further evidence was obtained by analyzing total DNA (withoutpre-digestion, see Materials and Methods) from wild type ordnd&Tsl infected cells bearing plasmid pl958 (20 and 30 min afterinfection). Such DNA was digested with AflU, AflR-EcoRI, PvuE-EcoRl or Mndlll and analyzed by Southern blot using either cat(plasmid) or pac (phage) probes. No AflR, but multiplerecognition sites for EcoRl and HindWl are present in the SPP1DNA; whereas pl958 is cut only once with A/IB, and PvuU anddoes not have any recognition site for EcoRl and HindUl. Sincecleavage at the SPP1 pac site occurs about 12 min after infection(our unpublished results), the experiments were designed withthe following criteria in mind: i) the stretch of homology betweenthe plasmid and the phage is 2.2 kb away from the phage pacsite (30, Fig. 6C), and ii) the restriction enzymes and specificradiolabeled DNA probes chosen would allow us to analyzewhether the putative phage: :plasmid chimeras shown in Fig. 6C



are generated. Two possible crossover products would result,one with phage sequences to the right of the homologous region(I) and the other with phage sequences to the left (II). It shouldbe noted that, only chimera I carrying a pac site would be thesubstrate for plasmid packaging.

When total DNA from wild type infected cells bearing plasmidpl958 was digested with AfttL-EcoRl or PvuR-EcoRl andanalyzed with the cat probe (Fig. 6A), an expected signal withthe plasmid linear form (2.5 Kb) and a weak signal to a 1.9 Kbfragment (lane l, AflFL-EcoKI) or 3.0 Kb fragment (lane 2, PvwII-EcoRT) were detected (see chimera II). At present, we are unableto explain the signal ofa 1.7 kb fragment which appears whentotal DNA is digested with PvuU-EcoRI (Fig. 6A, lane 2).

When total DNA was digested with AflH and analyzed usingthe pac probe, an expected signal with undigested SPP1 DNAand a weak signal corresponding to a 3.8 kb fragment weredetected (Fig. 6B, lane 1 and 2). This DNA fragment, whichis detected as early as 20 min after infection of wild type (Fig.6B, lane 1) or dnaB(Ts) cells (data not shown), might correspondto the AflU-pac generated fragment (see chimera I). That this3.8 kb pac-cnded DNA fragment, after Aflll digestion, is ofpl958 origin was confirmed by hybridization with the plasmidprobe (data not shown). Digestion of total DNA of infected cellsbearing plasmid pl958 with ffindHI gave rise to the expectedSPP1 HindUl fragments 1 and 6 (30) plus a band that co-migrateswith undigested DNA, when the pac probe was used (Fig. 6B,lane 3 to 6). The Mndlll resistant material accumulated as earlyas 20 min after infection seems to be independent of the strainin which the phage was amplified (wild type or dnaWil strain).Such material could represent concatemeric plasmid moleculescarrying the SPP1 pac site (chimera I).

From those data we conclude that: i) the formation ofconcatemeric plasmid::phage chimeras occurs as early as 20 minafter infection, and ii) both types of chimeras (I and II) areaccumulated and their rate of accumulation is independent of thednaB function. Since SPP1 packages its DNA unidirectionallyfrom pac (30, 31), we assume that only chimera I is the substratefor concatemeric plasmid DNA encapsidation.

DISCUSSION

The experiments presented above demonstrate that: i)concatemeric plasmid DNA is de novo synthesized followingSPP1 infection, ii) plasmid concatemeric replication is an activeprocess and not simply the result of passive replication uponplasmid integration into the phage genome as previously reportedin other host-phage systems (32, 33), and iii) phage: :plasmidchimeras carrying the pac site are generated as a consequenceof recombination-dependent concatemeric plasmid DNAsynthesis, and iv) the rate-limiting event defining the transductionfrequency is not the amount of concatemeric DNA generated ina phage infected cell.

Concatemeric plasmid DNA synthesis initiates simultaneouslywith the onset of phage replication (23). We could infer that twooverlapping modes of concatemeric plasmid DNA replicationexist. In the absence of homology between phage and plasmidDNAs, only inactivation of the ExoV enzyme would be requiredfor concatemeric plasmid DNA synthesis. This is logical sincesuch inactivation is needed for initiation of concatemeric phageDNA synthesis (18-20).

In the presence of homology, concatemeric plasmid DNA isdetected upon thermal inactivation of the dnaB gene product orin the recE polA genetic background (hdp DNA synthesis).

In E. coli, B. subtilis and 5. aureus the synthesis ofconcatemeric plasmid DNA is independent of plasmid encodedreplication functions (6, 9, 14, 15, 17, 34). In contrast, plasmidreplication was required for the synthesis of concatemeric plasmidmolecules in the 5. aureus phage 011 -pT 181 system (7).

The DNA to be packaged into SPP1 capsids is recognized andcleaved by phage-encoded proteins at a target sequence termedpac site (35, 31). The substrate, consisting of tandemly repeatedunits arranged in a head-to-tail configuration, is packaged by anunidirectional and processive 'headful' mode (29, 35, 31). Since(a) homology between phage and plasmid enhances the plasmidtransduction frequency (2, 6, 7, 9) and is a prerequisite for hdpDNA synthesis, (b) the appearance of a plasmid::phage chimeracarrying the phage pac site is detected as early as 20 min afterinfection, and (c) such chimera is accumulated under conditionsin which hip DNA synthesis is markedly reduced, we favour thatmodel 2 (see Introduction) is involved in the synthesis of packableconcatemeric plasmid DNA. Alternatively, the formation of thechimera could be a post-synthetic process, due to homologousrecombination between plasmid and phage concatemericmolecules (7).

ACKNOWLEDGEMENTS

We are very grateful to T.A. Trautner for discussions duringthe preparation of the manuscript. We are indebted to J.-F. Viretfor the communication of his unpublished results, and A.C. Stiegeand G. Luder for excellent technical assistance. This researchwas partially supported by Deutsche Forschungsgemeinschaft (Al284/1-1). A.B. was supported by an EMBO long term fellowship.

REFERENCES1. Low, K.B., and Porter, D.D. (1978) Ann. Rev. Genet. 12, 249-287.2. Lofdahl, S., Sjostrom, J. and Philipson, L. (1981) J. Virol. 37, 795-801.3. Canosi, U., Luder, G. and Trautner, T.A. (1982) J. Virol. 44, 431 -436.4. Takahashi, H. and Saito, H. (1982) Mol. Gen. Genet. 186, 497-500.5. Schmidt, C. and Schmieger, H. (1984) Mol. Gen. Genet. 196, 123-128.6. Alonso, J.C., Luder, G. and Trautner, T.A. (1986) EMBO J. 5, 3723-3728.7. Novick, R.P., Edelman, I. and Lofdahl, S. (1986) J. Mol. Biol. 192,

209-220.8. Liebeschuetz, J. and Ritchie, D.A. (1986) J. Mol. Biol. 192, 681-692.9. Kreuzer, K.N., Yap, W.Y., Menkens, A.E. and Engman, H.W. (1988) J.

Biol. Chem. 263, 11366-11373.10. Luder, A. and Mosig, G. (1982) Proc. Natl. Acad. Sci. USA 79, 101 -105.11. Kozinski, A.W. (1983) in Bacteriophage T4 (Mathews, C , (Cutter, E., Mosig,

G. and Berget, P. eds) pp 111-119, American Society of Microbiology,Washington.

12. Telander-Muskavitch, K.M. and Linn, S. (1981) in The Enzymes (Boyer,P.D., ed) Vol. 14, pp 233-250, Academic Press, New York.

13. Cohen, A. and Clark, A.J. (1986) J. Bacteriol. 167, 327-335.14. Viret, J.-F. and Alonso, J.C. (1987) Nucl. Acids Res. 15, 6349-6367.15. Silberstein, Z. and Cohen, A. (1987) J. Bacteriol. 169, 3131-3137.16. Viret, J.-F. and Alonso, J.C. (1988) Nucl. Acids Res. 16, 4389-4406.17. Leonhardt, H. (1989) Dissertation, Freien Universitat Berlin.18. Sakaki, Y. (1974) J. Virol. 14, 1611-1612.19. linger, R.C. and Clark, A.J. (1972) J. Mol. Biol. 70, 539-548.20. van Vliet, F. Couturier, M. de Lafonteyne, J. Jedlicki, E. (1978) Mol. Gen.

Genet. 164, 109-112.21. Alonso, J .C, Stiege, A.C. Tailor, R.H. and Viret, J.-F. (1988) Mol. Gen.

Genet. 214, 482-489.22. Ceglowski, P., Luder, G. and Alonso, J.C. (1990) submitted.23. Burger, K.J., and Trautner, T.A. (1978) Mol. Gen. Genet. 166, 277-285.24. Deichelbohrer, I. Alonso, J.C. Luder,G . and Trautner, T.A. (1985) J.

Bacteriol. 162, 1238-1243.25. Rottlander, E. and Trautner, T.A. (1970) Mol. Gen. Genet./OS, 4/-bO.26. Maniatis, T., Fritsch, E.F. and Sambrook, J. (1982) Cold Spring Harbor

Laboratory, Cold Spring Harbor, New York27. Alonso, J.C. and Trautner, T.A. (1985) Mol. Gen. Genet. 198, 427-431.



28. Mendelson, N., and Gross, S.D. (1967) J. Bacteriol. 94, 1603-1608.29. Casjens, S.R. (1985) In Virus Structure and Assembly (Casjens, S.R. ed.)

pp. 75-147, Jones and Bartlett Publishers Inc.30. Behrens, B., Luder, G., Behncke, M., Trautner, T.A., and Ganesan, A.T.

(1979) Mol. Gen. Genet. 175, 351-357.31. Bravo, A., Alonso, J.C. and Trautner, T.A. (1990) Nucl. Acids Res. 10,

2881-2886.32. Orbach, M. and Jackson, E.N. (1982) J. Bacteriol. 149, 985-994.33. Mattson, T., VanHouwe, G., Bolle, A. and Epstein, R. (1983) J. Mol. Biol.

170, 343-355.34. Novick, P.R (1989) Ann. Rev. Microbiol. 43, 537-565.35. Deichelbohrer, I., Messer, W. and Trautner, T.A. (1982) J. Virol. 42,

83-90.


the generation of concatemeric plasmid dna in bacillus subtilis as

Documents