View metadata, citation and similar papers at core.ac.uk brought to you by CORE

provided by Elsevier - Publisher Connector

Virology 468-470 (2014) 274–282

Contents lists available at ScienceDirect

Virology

journal homepage: www.elsevier.com/locate/yviro

Bacteriophage P1 pac sites inserted into the chromosome greatly increase packaging and of genomic DNA

Haomin Huang 1, Millicent Masters n

Institute of Cell Biology, University of Edinburgh, Kings' Buildings, Edinburgh EH9 3JR, UK

article info abstract

Article history: The Escherichia coli P1 packages host chromosome separately from phage DNA, and Received 5 March 2014 transfers it to recipient cells at low frequency in a process called generalized transduction. Phage Returned to author for revisions genomes are packaged from concatemers beginning at a specific site, pac. To increase transduction rate, 29 March 2014 we have inserted pac into the chromosome at up to five equally spaced positions; at least this many are Accepted 17 July 2014 fully tolerated in the absence of P1 infection. A single chromosomal pac greatly increases transduction of Available online 12 September 2014 downstream markers without decreasing phage yields; 3.5 as much total chromosomal DNA is Keywords: packaged. Additional insertions decrease phage yield by 490% and also decrease phage DNA synthesis, Phage P1 although less dramatically. Packaging of chromosomal markers near to and downstream of each inserted Pac site pac site is, at the same time, increased by greater than 10 fold. Transduction of markers near an inserted Packaging pac site can be increased by over 1000-fold, potentially allowing identification of such transductants by E. coli transduction E.coli chromosome engineering screening. & 2014 Elsevier Inc. All rights reserved.

Introduction The pac site is located on the 650 base-pair P1 EcoRI fragment 20 (Bächi and Arber, 1977). A fully functional 161 bp pac site Bacteriophage P1 is a temperate DNA coliphage with a double- contains four hexanucleotide “TGATCA/G” repeats at one end and stranded, linear, cyclically permuted, terminally redundant chro- three at the other, flanking a 13 bp sequence in which Pacase cuts. mosome (Lehnherr, 2005, Yarmolinsky and Sternberg, 1988). The Each repeat contains a GATC Dam methylation site, and methyla- full P1 genome sequence has been published (Lobocka et al., 2004; tion appears essential for recognition and cleavage by Pacases Yarmolinsky, 2004). The phage DNA consists of the 94 kb long (Sternberg and Coulby, 1990). The unpackaged end of the cut pac genome plus a 10–15 kb variable terminal redundancy, together site is degraded by host nucleases (Sternberg and Coulby, 1987a). equivalent to 2.2 Escherichia coli map units (minutes). The max- To prevent understuffing of the phage prohead and to ensure imum distance between cotransduced host DNA markers is thus circular permutation, the cleavage of phage concatemers at more limited to this length. than one pac site must be prevented. Although hypotheses have The phage chromosome is circularized by recombination early been offered (Yarmolinsky and Sternberg, 1988; Skorupski et al., after infection allowing the later formation of phage concatemers 1994a) the actual mechanism that prevents additional Pacase by rolling-circle replication (Cohen, 1983; Bornhoeft and Stodolsky, cleavages within concatemers is unknown. 1981). Concatemers are cut once by the phage-encoded enzyme, P1's widespread familiarity and usefulness results from its Pacase, at a specificsite,pac (packaging), from which the filling of ability to mediate generalized transduction (Masters, 1985, 1996, phage heads is initiated (Sternberg, 1990 (review), Skorupski et al., 2000), but the mechanism by which transducing phages arise is 1994, Skorupski et al., 1992, Sternberg and Coulby, 1987a, 1987b). not fully understood. Since transducing particles contain only host Sequential encapsidation from this cut yields the set of redundant, DNA (Ikeda and Tomizawa, 1965), which lacks pac sites (Hanks, permuted molecules found in progeny phages, a process termed 1986), and packaging of host sequences is mainly proportional to “headful” packaging. Each concatemer is thought to yield three to their intracellular concentrations (Masters, 1970, 1977, 1985; five infectious progeny (Sternberg and Maurer, 1991, Skorupski et Masters et al., 1984), packaging must also be able to begin from al., 1992). free host DNA ends. Such ends are found at replication forks or could result from nuclease action. Sequential packaging from pac-free chromosomes has been reported (Harriman, 1972; n Corresponding author. Tel.: þ44 131 650 5355; fax: þ44 131 650 8650. Hanks et al., 1988; Sternberg and Coulby, 1987b) and it appears E-mail addresses: [email protected] (H. Huang), – [email protected] (M. Masters). to proceed much farther than the 3 5 headfuls packaged from 1 Oncobiologics, Inc., 1 Duncan Drive, Cranbury, NJ 08512, USA phage concatemers. When transducing phage are prepared from

http://dx.doi.org/10.1016/j.virol.2014.07.029 0042-6822/& 2014 Elsevier Inc. All rights reserved. H. Huang, M. Masters / Virology 468-470 (2014) 274–282 275 cells in which a single pac site has been inserted in the chromo- the way in which these multi-pac strains behave as hosts for P1 some, processivity from pac can be inferred, confirming that infection and as transductional donors. Pacase does not need to cut at any but the initial pac site. The number of fully transduced bacterial progeny recovered for any particular marker is normally between 105 and 106 per Methods and materials phage even though transducing phages have been reported to comprise up to 6% of particles in lysates of P1kc (Hanks et al., Bacteria, , and associated techniques 1988). This is thought to be because most transductions are incomplete (abortive), terminating with a non-replicating, unili- Bacteriophage P1kc, purified from our laboratory stock, was nearly inherited stable circle in the recipient cell. Sandri and used throughout this work. All pac-containing strains were Berger, (1980a, 1980b) presented evidence that the circular DNA derived from E. coli K12 MG1655 obtained from F. Blattner. in abortive tranductants is not covalently closed and that circular- pKO3 (Link et al., 1997) was kindly provided by G. Church. Media ity is maintained by a bound protein. Unfortunately, the putative and methods for bacterial growth and transductional selections protein has not been identified, nor has abortive P1 transduction were as described by Masters (1970). Transductional recipients been studied since 1980. were MM303 (Masters et al., 1984), MM38 [MM18 (Jenkins et al., þ Our aim in this work was to increase the total number and 1986) asnB ], AB352 (thr-1 leuB6 lacZ4 purF1 rpsL8 thi-1 supE44); proportion of transducing phages in lysates. Although P1 mutants from CGSC, Yale University, and MG1655Δlac (this lab). Other have been isolated that transduce with increased frequency (Wall plasmids are described in Table 1. and Harriman, 1974; Iida et al., 1987, 1998), we were not able to exploit this (the Wall and Harriman mutant is no longer available Strain construction and the Iida mutants did not increase transduction in our system). Since single pac sites inserted into the chromosome greatly The argB〈〉pac strain, MG(argB〈〉pac), was made using the Link increase the frequency with which markers to one side of them et al., (1997) pKO3 procedure. Crossover PCR was used to construct are transduced (see above), and since it is now possible to a-BamH1-argC-sacI-argH-BamH1- fragment that was cloned into introduce exogenous DNA into the chromosome at any site with pKO3 to yield pHM2. P1 EcoRI-20 (containing pac) was cloned into relative ease (Link et al., 1997; Datsenko and Wanner, 2000), we the pHM2 SacI site to give pHM3, which was then used to replace inserted multiple pac sites into the chromosome hoping to argB with pac by transformation followed by temperature and increase transduction frequency proportionately. We report here sucrose resistance selection/screen. The four other single pac

Table 1 Principal plasmids used in this work.

Name Description Source

pHM3 pKO3 derivative containing argCH gene crossover PCR fragment and the pac site fragment cut from pPAC This work pPAC pBR325 Cm:: P1EcoR1.20 Hanks et. al., 1988 pCP20 pSC101-based vector expressing the Flp recombinase; repA(Ts), AmRKmR Cherepanov and Wackernagel, 1995 pTOF24 pKO3 ːː [HincII-HincII: 1252 bp, aph gene from pUC4K]; CmR KmR Ts SucS Merlin et al., 2002 pHM pTOF24 Δ[PstI-SalI: 1293 bp], ːː[PstI-SalI: polylinker, PstI-BamHI-PacI-SacI-SmaI-SwaI-NheI-NotI-AscI-SalI]; CmR Ts SucS This work

holC (96.6) 100/0 yjfJ (95) nadR 99.7) MetI (4.8) pgi (91.2) (3.0) phoA (8.6) pac[phoR-brnQ(pb)] 9.0min argB<>pac 89.5min ybaV (10) lacZ 7.8min squU (87.7) proC 8.7min ubiF (15) hemD (85.9) metB 88.9min (8) ilv 85min hcr (19.6) waaS (82) 81.7min yhjY (80) ptsG (24.9) zntA (77) MG1655 75 (showing pac sites) 25 yhfX (75.6) pyrF 28.9min trpA(C) (28.3) argG 71.5min rrlD (73.8) pac[dacB-yhbZ(dy)] 71.8min pac[tyrR-tpX(tt)] 29.9min dacB (71.7) purF 52.3min aroC 52.7min his 45min ssnA (65) rspR (35) 65min hycB (61.4)

hemF (55) yeeA (44.8) pac[dsdA-emrY(de)] 53.4min 50

intS (53.1) rhmR (50.9)

Fig. 1. Map of MG1655 showing locations of pac sites and markers used. The 5 pac site locations are shown as triangles with the apex on the circle indicating direction of packaging; rectangles show chromosomal positions of insertion sites (between indicated genes, in red) and mnemonics. Chromosomal positions (min.) of genes used as hybridization probes are indicated outside the ring (positions in blue). Genes inside the ring (turquoise) were used as transductional markers. metB and trpA were used for both purposes. Gene names and positions have been updated from EcoGene 3.0 (http://www.ecogene.org//). 276 H. Huang, M. Masters / Virology 468-470 (2014) 274–282

Table 2 P1 single cycle liquid lysis with synchronous infection Genes probed and insertion positions of pac sitesa. Overnight cultures were diluted 50-fold into L-Broth (LB) Marker/ pac site Location on Length EcoRV Orientation 3 2 þ 1 names chromosome fragment of pac containing 2.5 10 MCa and grown at 37 C with shaking 3 (min) (probe template) to an OD600 of 0.3–0.4. KCN was added at 2 10 M followed by phage P1 at an m.o.i. of 4. After 10 min, 5 ml aliquots were washed nadR 99.7 3.0 k with LB to remove KCN and Ca2 þ , diluted 10-fold into Ca2 þ -free LB holC 96.6 0.7 k and resuspended. OD was followed and samples taken as yjfJ 95.0 1.2 k 600 pgi 91.2 4.6 k required until lysis. To separate pellet and supernatant, aliquots pac(argB〈〉pac) 89 ↓ were cooled on ice and spun at 16,000 g for 15 min. in a micro- metB 88.9 8.0 k centrifuge. Pellets were washed with 4% NaCl, 0.1% Tween and squU 87.7 2.4 k 20 ng/ml Cm. Supernatants and pellets were resuspended in buffer hemD 85.9 2.0 k waaS 82.0 1.4 k and chloroform-treated. yhjY 80.0 1.7 k zntA 77.7 10.8 k Molecular techniques yhfX 75.6 3.0 k rrlD 73.8 7.0 k pac(dacB-yhbZ) 71.75 ↓ DNA was extracted from P1-infected cells as described by -dy Sternberg and Colby (1987a). For Southern blotting P1 DNA was dacB 71.7 3.8 k extracted from purified concentrated phage lysates, essentially as ssnA 65.0 2.5 k described by Hanks et al. (1988),digestedwithEcoRV (except for hycB 61.4 3.1 k hemF 55.0 13.5 k the experiment described in Fig. 3 where EcoR1 was used) and pac(dsdA-emrY) 53.4 ↓ fractionated on 0.8% agarose gels at 30 V overnight. DNA in gels was -de depurinated, alkali-treated, and the single-stranded DNA was trans- intS 53.2 4.7 k ferred to membrane filters by capillarity and then fixed with UV. rhmR 50.9 1.4 k Genes to be probed were PCR-amplified from chromosomal yeeA 44.8 1.8 k fi rspR 35.1 2.2 k DNA and puri ed using standard techniques. Probes for those to pac(tyrR-tpX) -tt 29.9 ↓ be used in a single experiment were made together by mixing the trpA(trpC) 28.3 5.9 k purified gene DNAs with the Roche DNA High Prime labeling kit ptsG 24.9 6.7 k and α 32P[dCTP]. High Prime mixture contains random oligonu- hcr 19.6 2.3 k ubiF 15.0 4.6 k cleotides, primers, Klenow polymerase and three dNTPs. Labeled ybaV 10.0 1.8 k probes generated are on average 80–120 bp in length, irrespective pac(phoR-brnQ) 9.0 ↓ of the size of template DNA. The Amersham Nick kit was used to -pb reduce the background of unincorporated radioactivity. phoA 8.6 9.7 k Denatured probes were hybridized to the filters and scanned metI 4.8 1.2 k with a Molecular Dynamics Phosphoimager. The densities of bands a Markers in plain type are contained within EcoRV fragments (lengths listed in were measured as pixel areas, using Imagequant software. P1 DNA column 3). Pac insertion positions, in bold type, are between the genes in brackets. band densities were proportional to DNA concentration, as fi Two-letter designations are used in strain names and on gures. expected. Digested chromosomal DNA was included on gels to enable calculations of relative probe labeling and the quality of hybridization. strains MG(1-pb), (1-de), (1-dy), (1-tt) (see Fig. 1 and Table 2 for In the experiment to measure total packaged DNA (Fig. 6) total locations and flanking genes) were constructed using pHH-1, a chromosomal DNA was used as the template with the High Prime pKO3 derivative based on those constructed by Merlin et al. Kit in order to produce labeled random fragments to use as probes. (2002). pHH-1 contains 8 restriction sites for the cloning of DNA The total label hybridized to the DNA in each column was flanking chosen intergenic pac insertion sites, the P1 pac fragment, recorded. Although label in chromosomal ladders was propor- R and a selection marker (aph, specifying Kan ). aph is flanked by tional to DNA loaded, this was not necessarily true of summed FRT sites to permit its later removal [using pCP20 (Cherepanov and transduced DNA bands. To render the intensities of each P1 DNA Wackernagel, 1995) to express Flp recombinase]. KanR recombi- sample proportional to volume, a marked intercept, calculated R nants were screened for loss of Cm and were verified by PCR. algebraically, and differing amongst the three DNAs, was sub- Multi-pac strains were constructed by successive P1 transductions tracted. Also subtracted is a [DNA]-adjusted fraction of non-pac R and removal of the Kan marker. site dependent packaged chromosomal DNA calculated from the R A single 2pac strain was made by transducing the Kan (1-pb) P1 chromosomal pac-free lane intensity. replacement to the argB〈〉pac strain. All strains containing more than 2 pac sites are derived from this MG1655(2pac) strain. Strains are named with reference to the number and identity of pac Results and discussion insertions. The number in the brackets is the number of pac replacements in the strain. MG (argB〈〉pac), the strain in which Packaging from a single pac site on the chromosome argB has been replaced by pac, does not have a numerical designation as it was constructed differently than the others. In order to determine the effect of a chromosomal copy of pac MGΔargB is a control strain from which argB has been deleted on transduction we inserted the pac-containing P1 EcoRI-20 and not replaced by pac. The other four strains with one pac site fragment (Bachi and Arber, 1977) into the chromosome in place are identified by the first letters of the genes between which pac of the argB gene (89.5 min). We chose argB for replacement as has been inserted (see Fig. 1, Table 2). The strains with 3 pac sites strains mutated in nearby genes were available, facilitating later are identified by their single unique site. MG(3-tt) is the parent of experiments. MG(4pac) to which the (dy) construct has been added; (5pac) We first confirmed that the pac insert increases the transduc- contains (de) in addition. tion frequency of a downstream chromosomal marker. The metB H. Huang, M. Masters / Virology 468-470 (2014) 274–282 277

P1( argB) P1(argB<>pac) MG1655 DNA

zntA (77.7) metB (88.9) rrlD (73.8) trpA (28.3) pgi (91.2) dacB (71.7) yhfX (75.6) yihU (87.7) hemD (85.6) yhjY (80) rfaS (81.9) yjfJ (95) 6 g 4 g 7.9 g5.3 g2.7 g 0.5 g1 g2g4g8g

P1 P1 P1 P1 ( argB)(argB<>pac) MG1655 DNA ( argB)(argB<>pac) MG1655 DNA

hemF (55) metB (89) metB (88.9) pstG (25) trp+ynj (28) ubiE (15) dacB (72) dacB (71.7) nadR (100) hycB (61.4) ssnA (65) hcr (20) ydfH (35) ybaV (10) yeeA (44.8) metI (4.8) rhmR (50.9) 4 g 5.3 g 4 g 2 g 4 g 5.3 g 4µg 2µg ) 100 argB

10

1

0.1 pac

Hybridization P1(argB<>pac)/ P1( 020406080100 Chromosomal units (min)

Fig. 2. Hybridization to transducing DNA from a single pac donor. (A) Chromosomal DNA and DNA extracted from three pairs [(MG1655(argB〈〉pac) and MG1655(ΔargB)] of separately prepared P1 lysates were EcoRV-cut, processed and hybridized to probe mixtures. Bands are identified with the names of probed genes followed by their map positions. (B) Data from 5 experiments (each with a different symbol and different probe mixes), are plotted against chromosomal positions. Ordinate values are band densities of P1(argB〈〉pac) relative to P1(Δarg), each normalized to trp hybridization. gene, at 88.9 min (Fig. 1, Table 2), is about 0.6 min. from argB; trpA To demonstrate enhanced packaging directly we analyzed three at 28.3 min is not expected to be packaged from this pac site. separately prepared pairs of EcoRV cut P1 lysates (7pac)using Surprisingly, argB〈〉pac increased the transduction frequency of Southern hybridization. Filters were hybridized to combinations of metB by as much as 1000 relative to the pac-free strain, whereas probes for 27 markers from around the chromosome (Fig. 2A, Table 2). the transduction frequencies of trpA from pac þ and pac-free We verified that band density was directly proportional to DNA donors were similar (ratio of 0.92). These data are included in concentration using the densities of bands in the chromosomal and Fig. 4B, below. P1(ΔargB)DNAladdersinthetopgelinFig. 2A. Values were normal- ized by setting Trp þ hybridization from MG1655ΔargB lysates¼1. The P1 packaging of MG1655 (argB〈〉pac) DNA ratios of P1(argB〈〉pac)toP1(ΔargB) hybridization were calculated and plotted against marker position (Fig. 2B). Increased packaging of chromosomal DNA downstream from pac In Fig. 2B, the data from 5 independent hybridization experi- has previously only been inferred from increases in transduction ments (including those in Fig. 2A) are shown. The ratio of frequency (Sternberg and Coulby 1987a, Hanks, 1986, PhD Thesis). hybridization to P1(argB〈〉pac) to that of P1(ΔargB) was calculated 278 H. Huang, M. Masters / Virology 468-470 (2014) 274–282 for markers distributed over the chromosome and is plotted Although quantifying abortive transduction could be informative, against marker position. Clearly, the DNA encoding markers to we could find no simple and reproducible way to do so. one side of pac is more highly represented within P1(argB〈〉pac) than within P1(ΔargB) DNA and packaging is increased by as much A single pac at other chromosomal locations as 30-fold for markers closest to the inserted pac. This decreases rapidly over 20 map units, but some increase in packaging is To introduce pac sites at other locations on the chromosome, detectable as far as 35 min downstream from the site of packaging four additional single pac-containing strains were made by insert- initiation. Packaging from the chromosomal pac site is clearly ing pac in the same orientation as (argB〈〉pac) between genes in directional, since the packaging of markers to only one side of pac four new positions at 20 min intervals (Table 2 and Fig. 1). is increased. Packaging of markers to the other side appears P1 transduction was again used to test whether transduction markedly reduced, consistent with the finding that the packaged frequency from the new strains with single pac sites was stimu- DNA end is protected from cellular nucleases, while the other end lated to one side of pac (Fig. 4B, below). A marker within 2 min of is degraded (Sternberg and Coulby, 1987a). The reduction in pac was tested for three of these, (the nearest marker then packaging we see between 91 and 100 min (Fig. 2B) could be available for (1-de) was 8 min from pac). Transduction was again accounted for by such degradation. Markers on about half of the greatly increased [up to 1300-fold for ArgGþ from MG1655(1-dy)]. chromosome (from 50 min counterclockwise (passing trp,to In Fig. 4B (below) transduction frequency is plotted vs. distance 91 min) are packaged between 90% and 50% as efficiently as from from the nearest pac site with the leftmost five Xs representing P1(ΔargB). This may be because the number of phage heads within markers packaged early from the five single-pac strains. We an infected cell is limiting and packaging from bona-fide pac sites conclude that the effect of pac on packaging of transducing DNA favored over those originating from random cuts. is independent of the location of pac on the chromosome. We noted above that metB transduction can be increased 1000- fold by contiguity to pac; how this occurs despite an only 15-fold increase in actual packaging is unknown but it is also observed for Lysate preparation from strains with multiple pac sites other markers close to a pac site (see below). One possibility is that many likely abortive transductants are now capable of undergoing Strains with multiple pac sites were constructed as described recombination, perhaps because there is not enough of the above. P1 plate lysates were then prepared on MG1655, the five putative protein presumed to maintain circular DNA available for single-pac containing strains, the 2-pac strain, three strains, each the 15-fold increased amount of packaged host DNA (see below). with a different 3rd pac, and the 4-pac and 5-pac strains.

1.0 1.0 5pac 600 600

0.1 . O.D. O.D. MG P1, KCN Wash 0.4 0.01 01020 30 40 50 0 50 100 150 200 250 1000 Minutes of growth 1010 MG 9 10 5pac 108 107 16.5 kb

106 P1 DNA/ cell 2.8kb

Titer 100 105 104 103 Minutes after dilution 102 160 180 200 220 240 µ Minutes of Growth MB1 gP1 DNA MG1655 (5pac) MG1655 4 2 1 40” 30” 20”10” 0” 30” 20” 10” 0” 16.5 kb 10kb

3kb 2.8kb 2kb

Fig. 3. Effect of multiple donor pac sites on lysis by P1. (A) MG1655(B), MG1655(argB〈〉pac)(K), MG(5pac)(’), and MG(3-tt) (Δ) were grown in LB to an O.D.600 of 0.3 and P1 added at an moi of 4 in the presence of cyanide. After 10 min. cultures were washed, resuspended in LB and followed until lysis. (B) P1 in the supernatants (closed symbols) and in the CHCl3-lysed pellets (open symbols) were measured by titration in samples taken after resuspension from MG1655(argB〈〉pac)(K) and MG(5pac)(’) cultures. (C) Growth of MG(5pac) and MG1655 in a similar experiment after P1 adsorption and culture dilution. (D) Samples were taken from (C) during phage development; intact cells were collected by centrifugation and retained for DNA analysis. DNA, purified from cells, was cut with EcoRI and hybridized to labeled probes detecting 16.5 kb and 2.8 kb P1 bands. Results shown are from analysis of the phosphoimage in (E). (E) Time course of labeling of P1 16.5 kb and 2.8 kb EcoRI fragments from which the time course of P1 DNA synthesis was calculated [plotted in (D)]. The extra band in the 4 μg P1 DNA track results from incomplete hydrolysis. The MB1 ladder shows sizing standards separated by 1 kb. H. Huang, M. Masters / Virology 468-470 (2014) 274–282 279

Pac sites 4: 3: 2: Pac: arg<>pac dy de tt pb 10000 10000 AB

donor metB 1000 lac X -free pac

argG 100 * his trp * ilv 10

1 Tranductants relativeto 1 100 90 80 70 60 50 40 30 20 10 0 0.1 1 10 100 Chromosome units (min) Distance from pac

Fig. 4. Tranduction from multi-pac donors. (A) Lysates prepared from MG1655 and 2-pac,3-pac and 4-pac derivatives were used to transduce six markers selected to be near the origins of packaging from each pac site (positions indicated above figure and by dashed vertical lines). Transductants per phage, relative to those from MG1655 as host, are plotted against chromosmal position in Fig. 5A. (B) The values from Fig. 5A, plus others obtained from the 5 single-pac strains and trpA (Xs), are plotted as a function of distance from the nearest upstream pac site. Note that the X-axis in Fig. 4B is exponential. *indicates Ilvþ transductants; see text.

Surprisingly, although MG1655 and each of the single pac chromosomal map position, shows that markers very near to pac strains yielded plate lysates of P1 with high titers of between sites are transduced with very high frequencies, 103–104 higher 5 1010 and 1011 pfu/ml, all multiple pac strain lysates had titers of than from MG1655. This frequency drops by at least 90% over 20% only 103 to 105 pfu/ml. In liquid broth, before infection, all strains of the chromosomal length [see metB to argG for MG1655 (2pac)] divided with about the same doubling time (Fig. 3A). Thus the and the same again over the next 20%. Subsequent pac sites restore multiple pac insertions must interfere with growth after infection a high level of transduction. If there is a limit to how many pac or with production and/or release of progeny phage. To investigate sites can be inserted before packaging efficiency near pac further, we followed phage development in single-cycle liquid decreases, it must be greater than four. From Fig. 4B, in which lysates of MG1655 and its (argB〈〉pac), (3-tt) and (5pac) derivatives. transductant numbers are plotted against distance from the Growth after infection was at the same rate for all strains (Fig. 3A); nearest upstream pac site, we can see that transductant numbers however, the pac-free and single-pac strains lysed sooner and decrease exponentially with distance from pac over a greater than more completely than the multi-pac strains, achieving titers 100 range, irrespective of which pac site is closest. greater than 109. The multi-pac strains only reached titers of Note that ilv, marked by n in Fig. 4, is transduced less efficiently o5 108, fortunately just sufficient for adequate transduction. than expected. Ilv transduction normally occurs with higher Fig. 3B shows the accumulation of free phage for the 5pac and pac- frequency than predicted from the proportion of ilv DNA in a free hosts. The 5pac host not only lyses more slowly and releases lysate (Hanks et al., 1988), a property shared by nearby markers fewer phage, but it also retains fewer viable phage in the pellet. To also able to be cotransduced with oriC. This may be because see whether DNA is synthesized in the multi-pac strains but not replication is initiated on abortively transduced fragments con- incorporated into phage heads, we measured P1 DNA accumula- taining oriC and the ends thus created stimulate recombination tion in chloroformed lysates of MG1655 and MG1655(5pac)by (Masters, 1977). Ilv is 0.6 min from oriC, so that most random hybridization. Fig. 3C shows a lysis curve, 3E an autoradiogram and 2.2 min P1 “headfuls” of DNA containing one also contain the 3D P1 DNA/cell. The 5pac strain lyses later and has less P1 DNA/cell other. However, in the construct argG〈〉pac, headfuls originating at than the pac-free strain (Fig. 3D and E). The difference in P1 DNA pac would not include both ilv and oriC, making a pac-free strain content (1.6-fold) is not so great as is the titer (7-fold), an inappropriate control for ilv transduction from an argB〈〉 suggesting that packaging of the DNA rather than its replication pac donor. is principally responsible for the reduced titer. This may be due either to an insufficiency of heads or a fault in the packaging Hybridization using DNA prepared from strains with multiple process. The 1000-fold reduced titers of the plate lysates most pac sites probably result from outgrowth of host lawn before sufficient phage are produced to infect all cells, thus drastically reducing the To determine whether competition between multiple pac sites number of cycles of infection and yield of phage. limits chromosomal packaging, DNA prepared from P1 grown on pac containing strains for a single cycle in broth were hybridized to Transduction using lysates prepared from strains with multiple pac several probes. sites Representative Southern blots (Fig. 5A) show relative hybridi- zation of markers from five strains with single, and four strains In order to determine whether multiple pac sites reduce the with multiple pac inserts. In Fig. 5B, band intensities relative to amount of packaging initiated at each pac site, single growth cycle those from a pac-free strain (MG1655) are plotted for six of these transducing lysates were prepared from MG1655 and its 2pac, (symbols above Fig. 5A). The multi-pac strains hybridize strongly 3-de and 4pac derivatives. These were used to transduce a set of to bands downstream of and close to inserted pac sites, as markers selected so that each is an indicator of packaging initiated expected. However, when compared with the hybridization of P1 from one of the five possible chromosomal pac site locations. (argB〈〉pac) to bands originating far from pac, such as trpAC [by P1 Fig. 4A, in which transduction ratios are plotted against (3-dy) or P1(3-de) lysates], yhfX,orholC, hybridization of multi- 280 H. Huang, M. Masters / Virology 468-470 (2014) 274–282

Fig. 5. Transducing DNA packaged from chromosomes with different pac site locations and numbers. (A) DNA from P1 lysates of MG1655 and nine derivatives with combinations of 0–4 pac sites was analyzed. Symbols above the relevant lanes are those used in (B) where the relative hybridization of 6 strains to 12 probes from around the chromosome is shown. Pac sites with positions (peaks) are named above the figure [(argB〈〉pac)¼arg]. (C) Relative hybridization of other lysates to markers between pac〈〉 argB and pac (1-dy) for 5 strains with (pac〈〉argB): 4pac (-&-, -Δ-), 3-dy (-B-), 3-tt (--), 2pac (-x-), pac〈〉argB (-m-) and 3 without: 1-tt (-J-), 1-pb (-’-), 1-dy (-~-). Results are corrected for DNA loading. Hybridizations are not shown.

contributes an EcoRV site converting the chromosomal EcoRV - MG1655 P1(5pac) P1(3-dy) P1(1-de) P1pac MBI fragment into two smaller ones.) 168 4 6 1.5 0.75 6 1.5 0.75 4 2 1 11µg DNA Processivity of packaging might decrease for reasons including KB chromosomal degradation from the unpackaged end of Pacase-cut 10 fi 8 DNA or because of insuf cient packaging capacity (which would 6 also reduce random initiation of packaging). To learn more DNA 5 extracted from P1 lysates grown on the (argB〈〉pac), (1-dy), (1-de), 4 (1-tt). (2pac), 2 (3pac) and the (4pac) strains were hybridized to 3.5 3 measure in more detail packaging between adjacent pac insertions 2.5 at 89.5 and 71.8 min (Fig. 5C, gel not shown). Six of the lysates 2 were from a donor with the argB〈〉pac insert. These show differ- ences in the rate of packaging decrease amongst them, with 1.5 packaging from the argB〈〉pac,2pac and 3-tt strains, which all have (argB〈〉pac), but not pac(1-dy), exhibiting higher transduction of markers closer to the “dy” replacement site, than do 3-dy or 4pac donors, which contain both inserts, consistent with possible degradation of the unprotected DNA end after a cut at pac(1-dy). 10 Repetition of the experiment with other pairs of pac sites would be 8 desirable to confirm the generality of these hybridization profiles. 6 5 4 Total chromosomal DNA packaged by P1 phage 3 2.5 Although markers to one side of pac are packaged at increased 2 frequency, reduced packaging far from pac could result in total chromosomal DNA packaged remaining the same. To measure the 1.5 total chromosomal DNA packaged, purified E. coli DNA was used as a template to make random probes covering its entire length. This should generate labeled probes for all chromosome sequences, Relative P32 13.9 3.6 1.7 4 1.9 1 3.2 allowing measurement of the total transducing DNA packaged by cpm per lane 11.7 2.8 1.5 P1 despite the presence of excess P1 DNA. P32 per µg DNA 2.3 2.4 2.3 2 1.9 2 1 0.95 1 0.3 In the gel shown in Fig. 6A, MG 1655 chromosomal fragments Fig. 6. Total chromosomal DNA packaged from donors with different chromosomal vary continuously in size, while in lysates only P1 bands are fi pac site numbers. Different amounts of each DNA con rmed that the signal visible. Fig. 6B shows the hybridization result. All lanes show obtained is proportional to loading. (A) P1 lysate and host DNA were cut with EcoRV and, together with size markers, were separated on gels. (B) DNA was bands of many sizes but more intense bands appear in each. A transferred to membranes, hybridized to 32P-labeled random chromosomal probes band between 2.5 and 3 kb appears to be shared by all lysates and and radioactivity/lane determined as explained in the text and the legend to Fig. 3. chromosomal DNA, suggesting that it was preferentially amplified Each lane is labeled with the amount of DNA applied. Labels below indicate relative during primer synthesis, as template DNAs should be equimolar. μ counts, normalized to 1 g of P1(1-de). The strongly hybridizing bands in the P1 lanes differ from those in pac DNAs seems reduced or absent (Fig. 5A). This suggests that the chromosomal lanes, are not coincident with P1 fragments, and processivity of packaging from each pac site is reduced in multi- are similar in the 3-dy and 5pac lanes, which share three pac pac donors. (Note that the dacB band is reduced in size in strains locations. The P1(1-de) DNA lacks the most intense of these (ca. with the (1-dy) pac insert; this is because the Pac insertion 8 kb in size), but displays others shared with 5pac (i.e. 4 kb). These H. Huang, M. Masters / Virology 468-470 (2014) 274–282 281 results are consistent with the hypothesis that the bands represent chromosomes with multiple pac sites can be packaged at the regions downstream from pac sites. expense of P1 DNA. The total amount of hybridizing DNA in each lane in Fig. 6Bwas determined. All values, expressed relative to that of P1(1-de), are Possible applications and future prospects listed below the lanes in Fig. 6B, and, per unit DNA loaded, below this. Using these measurements and those from a second gel (not Insertion of a pac site near to and in the appropriate orientation shown), we can deduce that pac sites on the chromosome increase could increase transduction sufficiently to allow transfer and chromosomal DNA packaging by P1. Inserting a single pac site into identification of an unselectable marker with minimal screening. the chromosome results in a 3.5-fold increase in the amount of The intriguing questions of whether and how abortive transduc- packaged transducing DNA. The total amount of DNA packaged tants may be stabilized remain to be addressed. from multi-pac strains increases with the number of pac sites but is not in direct proportion to pac number. Three inserted pac sites doubles rather than triples the amount of chromosomal DNA Acknowledgments packaged and five pac sites increases packaging only by a further 15–20%. Since packaging near each pac site, even when there are This work was supported by a Darwin Trust post-graduate five, is high, no packaging site appears to be used preferentially. scholarship to H. Huang, and is reported more fully in his PhD Possible explanations for the lack of direct proportionality Thesis (Edinburgh 2003). We would like to thank C. Merlin and S. between pac site numbers and DNA packaged include: (1) All or McAteer for discussion and technical assistance and W. Donachie many of the pac sites present are used in all cells which package for critical reading of the manuscript at each revision. transducing DNA, but there is a reduction in packaging of DNA remote from pac. (2) Only one or two pac sites in an References individual cell are used, but the increase in pac site number results in an increase in the percentage of cells that make transducing Bächi, B., Arber, W., 1977. Physical mapping of BglII, BamHI, EcoRI, HindIII and PstI phage. Distinguishing between these alternatives will require restriction fragments of bacteriophage P1 DNA. Mol. Gen. Genet. 153, 311–324. further experimentation. Bornhoeft, J.W., Stodolsky, M., 1981. Lytic cycle replicative forms of bacteriophages P1 and P1dl: concatemer forms. Virology 112, 518–528. Cohen, G., 1983. Electron microscopy study of early lytic replication forms of bacteriophage P1 DNA. Virology 131, 159–170. Conclusions Cherepanov, P.P., Wackernagel, W., 1995. Gene disruption in Escherichia coli:TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic- – Physiological effects of chromosomal pac sites resistance determinant. Gene 158, 9 14. Datsenko, K.A., Wanner, B.L., 2000. One-step inactivation of chromosomal genes in Escherichia coli K-12 using PCR products. Proc. Natl. Acad. Sci. U. S. A. 97, A single chromosomal pac site does not alter the course of 6640–6645. phage maturation or lysis. In liquid infections, two or more Hanks, M.C. (1986). PhD Thesis, Edinburgh University. Hanks, M.C., Newman, B., Oliver, I.R., Masters, M., 1988. Packaging of transducing chromosomal pac sites extend the period before lysis and reduce DNA by bacteriophage P1. Mol. Gen. Genet. 214, 523–532. viable phage yield by about 90%. Phage DNA synthesis is more Harriman, P.D., 1972. A single-burst analysis of the production of P1 infectious and normal, at 70% of the yield from hosts with pac-free chromosomes, transducing particles. Virology 48, 595–600. Iida, S., Hiestand-Nauer, R., Sandmeier, H., Lehnherr, H., Arber, W., 1998. Accessory suggesting a reduction primarily of packaging. Growth of all genes in the darA operon of bacteriophage P1 affect antirestriction function, strains is normal in the absence of P1 infection, even with five generalized transduction, head morphogenesis, and host cell lysis. Virology inserted pac sites. 251, 49–58. Iida, S., Streiff, M.B., Bickle, T.A., Arber, W., 1987. Two DNA antirestriction systems of bacteriophage P1, darA, and darB: characterization of darA– phages. Virology Packaging 157, 156–166. Ikeda, H., Tomizawa, J., 1965. Transducing fragments in generalized transduction by phage P1. Molecular origin of the fragments. J. Mol. Biol 14, 85–100. P1 packaging of markers to one side of and near to a Jenkins, A.J., March, J.B., Oliver, I.R., Masters, M., 1986. A DNA fragment containing chromosomal pac-site is increased more than 10-fold. This effect the groE genes can suppress mutations in the Escherichia coli dnaA gene. Mol. diminishes with distance but packaging can be detected over 30% Gen. Genet. 202, 446–454. Łobocka, M.B., Rose, D.J., Plunkett 3rd, G., Rusin, M.M., Samojedny, A.A., Lehnherr, of the length of the chromosome. One chromosomal pac site more H., Yarmolinsky, M.B., Blattner, F.R., 2004. Genome of bacteriophage P1. J. than triples the amount of chromosomal DNA packaged by P1. Bacteriol. 186, 7032–7068. Packaging of DNA is high downstream of each pac site of a multi- Lehnherr, H., 2005. Bacteriophage P1. In: Calendar, R. (Ed.), The Bacteriophages. – pac donor, but the total amount of DNA packaged does not Oxford University Press, New York, NY, pp. 350 361. Link, A.J., Phillips, D., Church, G.M., 1997. Methods for generating precise deletions increase proportionately with pac site number, perhaps because and insertions in the genome of wild-type Escherichia coli: application to open of increased degradation of donor DNA behind the pac cut. reading frame characterization. J. Bacteriol. 179, 6228–6237. Masters, M., 1970. Origin and direction of replication of the chromosome of E. coli B/r. Proc. Natl. Acad. Sci. U. S. A. 65, 601–608. Competition of transducing and phage DNA for encapsidation Masters, M., 1977. The frequency of P1 transduction of the genes of E.coli as a function of chromosomal position: preferential transduction of the oriC region. Mol. Gen. Genet. 155, 197–202. MG1655 (5pac) lysates contain about seven-fold more transdu- Masters, M., 1985. Generalized transduction. In: Scaife, J., Leach, D., Galizzi., A. cing DNA than those of MG1655, and about equally fewer (Eds.), Genetics of BacteriaAcademic Press, London197–215. infectious phage (Figs. 4 and 6). How might this come about? Masters, M., 1996. Generalized transduction. In: Neidhardt, F.C. (Ed.), Escherichia Encapsidation of P1 DNA requires cutting at pac by Pacase, which coli and Salmonella typhimurium: Cellular and Molecular Biology, 2nd edition ASM Press, Washington, DC, pp. 2431–2441 (Editor-in-Chief). is synthesized at a low level throughout infection (Skorupski et al., Masters, M., 2000. Transduction: host DNA transfer by bacteriophages. 2nd edition 1992). PacA binds to methylated or hemimethylated sites inde- In: Lederberg, J. (Ed.), The Encyclopedia of Microbiology, vol. 4. Academic Press, – pendently of PacB, but full methylation and PacB is required for San Diego, CA, pp. 637 650 (Ed-in-chief). Masters, M., Newman, B., Henry, C.M., 1984. Reduction of marker discrimination in cutting (Sternberg, 1990). Since Dam-methylated chromosomal transductional recombination. Mol. Gen. Genet. 196, 85–90. pac sites are present at the time of infection it is possible that Merlin, C., McAteer, S., Masters, M., 2002. Tools for characterization of Escherichia Pacase binds to them before fully methylated P1 concatemer is coli genes of unknown function. J. Bacteriol. 184, 4573–4581. Sandri, R.M., Berger, H., 1980a. Bacteriophage P1-mediated generalized transduc- synthesized. This would reduce the amount of Pacase available to tion in Escherichia coli: fate of transduced DNA in recþ and recA recipients. initiate phage DNA packaging, and could explain how DNA from Virology 106, 14–29. 282 H. Huang, M. Masters / Virology 468-470 (2014) 274–282

Sandri, R.M., Berger, H., 1980b. Bacteriophage P1-mediated generalized transduction in Sternberg, N., Coulby, J., 1987b. Recognition and cleavage of the bacteriophage P1 Escherichia coli: structure of abortively transduced DNA. Virology 106, 30–40. packaging site (pac). II. Functional limits of pac and location of pac cleavage Skorupski, K., Pierce, J.C., Sauer, B., Sternberg, N., 1992. Bacteriophage P1 genes termini. J. Mol. Biol. 194, 469–479. involved in the recognition and cleavage of the phage packaging site (pac). Sternberg, N., Coulby, J., 1990. Cleavage of the bacteriophage P1 packaging site (pac)is J. Mol. Biol. 223, 977–989. regulated by adenine methylation. Proc. Natl. Acad. Sci. U. S. A. 87, 8070–8074. Skorupski, K., Sauer, B., Sternberg, N., 1994. Faithful cleavage of the P1 packaging Sternberg, N.L., Maurer, R., 1991. Bacteriophage-mediated generalized transduction site (pac) requires two phage proteins, PacA and PacB, and two Escherichia coli in Escherichia coli and Salmonella typhimurium. Methods Enzymol. 204, 18–43. – proteins. IHF and HU. J. Mol. Biol. 243, 268 282. Wall, J.D., Harriman, P.D., 1974. Phage P1 mutants with altered transducing abilities Sternberg, N., 1990. Bacteriophage P1 DNA packaging. In: Adolph, K.W. (Ed.), for Escherichia coli. Virology 59, 532–544. Chromosomes: Eukaryotic, Prokaryotic, and Viral, vol. III. CRC Press, Boca Raton, Yarmolinsky, M.B., 2004. Bacteriophage P1 in retrospect and in prospect. J. – FL, pp. 225 240. Bacteriol. 186, 7025–7028. Sternberg, N., Coulby, J., 1987a. Recognition and cleavage of the bacteriophage P1 Yarmolinsky, M.B., Sternberg, N., 1988. Bacteriophage P1. In: Calendar, R. (Ed.), The packaging site (pac). I. Differential processing of the cleaved ends in vivo. J. Mol. Bacteriophages, vol. 1. Plenum Press, New York, NY, pp. 291–438. Biol. 194, 453–468.