Vol. 176, No. 11

Isolation and Characterization of a Generalized TransducingPhage for Xanthomonas campestris pv. campestris

B. D. WEISS,1 M. A. CAPAGE,1'2 M. KESSEL,"13 AND S. A. BENSON'*Department of Microbiology, University of Maryland, College Park, Maryland 207421; NeXagen, Inc.,

Boulder, Colorado 803012; and Department of Membrane and Ultrastructure Research,Hadassah Medical School, Jerusalem 91-010, Israel3

Received 28 December 1993/Accepted 19 March 1994

We have isolated and characterized a lytic double-stranded DNA Xanthomonas campestris pv. campestrisbacteriophage (XTP1) capable of mediating generalized transduction. The phage transduces chromosomalmarkers at frequencies of 10-5 to 10-6 transductants per PFU. We demonstrated its genetic utility by theisolation and cotransduction of linked transposon insertions to a nonselectable locus, xgl, required for thecleavage of 5-bromo-3-chloro-indoyl-,-D-galactoside and showed that rifand str alleles in X. campestris are 75%linked. One-step growth experiments showed that the latent and rise periods were each 2 h and the average

burst size was 35. The DNA genome is approximately 180 kb, presumably modified in a sequence-specificmanner, and may be covalently attached to protein(s). Electron micrographs show the phage particle to havean icosahedral head and contractile tail with tail fibers uniquely attached to a location 40 nm proximal fromthe end of the tail.

Xanthomonas campestris pv. campestris is a gram-negativeplant pathogen that causes black rot in crucifers (35). It is anagriculturally and industrially important organism that pro-

duces an exopolysaccharide (EPS) xanthan gum that maycontribute to the pathogenesis of X campestris (7, 30). Xan-than is an important food additive and a viscosifying agent ofcommercial interest in the oil and food industries (26).A major limitation in studying Xanthomonas species is that

there is no simple way to move mutations from one strain toanother, thus restricting the use of genetic approaches. Untilnow, the transfer of mutations has been limited to conjugalmating with broad-host-range plasmids (17, 22), a system thatrequires the lengthy processes of cloning and marker exchange.A simple means of transferring mutations between strainswould simplify genetic approaches with this organism. Inparticular, bacteriophage-mediated transduction would permitgenetic manipulations similar to those possible with othergram-negative bacteria with bacteriophages that allow gener-alized and specialized transduction. Bacteriophages capable ofinfectingX campestris have been reported (15, 16, 34), as havephages specific for related strains (5, 6, 8, 18, 32). However,none has been shown to mediate generalized transduction.We present here the isolation and characterization of the

first generalized transducing phage for X campestris pv.campestris. We have named this phage XTP1 (Xanthomonastransducing phage no. 1).

MATERIALS AND METHODS

Bacterial strains and media. Table 1 lists the bacterialstrains used. All X campestris pv. campestris strains werederived from strain B1459S-4L (S4L). Strain X1231 is a

rifampin-resistant (Rif) derivative of S4L containing a 16-kbchromosomal deletion that removes a cluster of genes (gum)

* Corresponding author. Mailing address: Department of Microbi-ology, University of Maryland, College Park, MD 207424. Phone: (301)405-5478. Fax: (301) 314-9489. Electronic mail address: sb77@umail.umd.edu.

required for xanthan biosynthesis (4). Strains BP302 andBP308 are nitrosoguanidine-induced mutants of strain X77.The xgl locus codes an enzyme that cleaves 5-bromo-3-chloro-indoyl-13-D-galactoside (X-Gal) but not o-nitrophenyl-galacto-side in vitro. BP108 was generated by transposon mutagenesis.All cultures were maintained on YM media (Difco) plusglucose (YMG). Liquid cultures were routinely grown inLuria-Bertani (LB) medium (27) at 30°C with aeration. Anti-biotics (,ug/ml) were added as follows: streptomycin (Str), 25;rifampin (Rif), 50; kanamycin (Kan), 30; tetracycline (Tet), 3.The substrate X-Gal was used at a concentration of 66 jig/ml.Antibiotics and other chemicals are from Sigma ChemicalCompany unless otherwise indicated. Restriction endonucle-ases were from Promega or New England Biolabs.

Mutagenesis. Mutagenesis was performed as described byMiller (19). A final concentration of 5 ,ug of nitrosoguanidineper ml was required to produce 50% killing of strain X77. TnJOand Tn9O3 insertion mutations were obtained by filter matingEscherichia coli HB101 carrying pRK2013(fQTn903)fQTn1O(10) to the recipient Xanthomonas strain. Plasmid pRK2013 isself transmissible and contains a ColEl replicon which isnonfunctional in Xanthomonas species. Transposon recipientswere selected on YM plates containing Rif and either Tet or

Kan. Transposon insertion pools were formed from approxi-mately 1,500 colonies by combining exconjugants from 15 to 30plates. The pools were stored at -70°C in 13% glycerol.

Transductions. Phage lysates were tested for transductionability by mixing successive dilutions of the phage stock (1/5,1/50, and 1/500) (0.1 ml each) with 0.4 ml of an overnightculture of recipient cells and incubating the mixture at 30°C for15 min. Following the incubation period, 1 ml of lx VogelBonner salts (VB) (33) was added. The cells were thencollected by centrifugation, resuspended in 0.1 ml of lx VB,spread onto YM plates, and incubated at 30°C for 2 h.Antibiotics were then added in a soft agar overlay. Transduc-tants appeared after 3 to 4 days. Transduction frequency isexpressed as transductions per input PFU.

Isolation and growth ofX. campestris phages. Soil samples ofapproximately 10 g each were suspended in 15 ml of LB broth

3354

JOURNAL OF BACTERIOLOGY, June 1994, P. 3354-33590021-9193/94/$04.00+0Copyright e 1994, American Society for Microbiology

on May 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

X CAMPESTRIS PHAGE 3355

TABLE 1. Xanthomonas strains used in this study

Strain Relevant characteristics Source and/or reference

X campestris pv. campestrisS4L Wild type Synergen, Inc., Boulder, Colo.BP109 nf-I Rif Spontaneous mutant; this studyX77 nf-2 Rif Synergen, Inc.X1231 Aguma::tet nf-2 Gum- Tetr Rif Synergen, Inc.BP108 X1231 (xgl-59b::Tn9O3 Kanr) Transposon mutagenesisBP1 12 str-9 Strr Spontaneous mutantBP125 str-9 nf-J Strr Rif Transduction; this studyBP302 xgl-4b nf-2 Rif Nitrosoguanidine mutagenesisBP308 xgl_85b nf-2 Rif Nitrosoguanidine mutagenesisBP353 BP302 (pig-1c::Tn1O Tetr) Transposon mutagenesis

X campestris pv. phaseoli XP6022 Strr Anne Vidaver, University of Nebraska, Lincoln

X campestris pv. pruni Edwin Civerolo, USDA, Beltsville, Md.

X campestris pv. vesicatonia Steve Hutcheson, University of Maryland, College Parka Construction of Agum is described in reference 4.b The designation xgl refers to a locus that specifies cleavage of X-Gal, resulting in a blue-green colony on media with this indicator.c The designation pig refers to a locus that confers yellow pigmentation to the Xanthomonas colony.

and permitted to shake for 30 min in a 20°C water bath. Soilsediment was removed by centrifugation, and the supernatantswere collected in sterile flasks. Log-phase cultures of X1231and S4L (25 p.l each) were added and allowed to growovernight at 30°C. The samples were treated with chloroform,clarified by centrifugation, and filtered through a GelmanSupor-450 0.45-pum-pore-size membrane filter. The filtrateswere assayed for Xanthomonas-specific phages by plating suit-able dilutions on strain X1231 or S4L in soft agar on LBmedium. Plaques were purified by dilution and used to makehigh-titer stocks by liquid lysis in LB broth and soft agaroverlay (27). Lysates were treated with chloroform for severalminutes, clarified by centrifugation, and stored at 4°C in theabsence of chloroform.Phage lysate production from transposon pools. One milli-

liter of the transposon pool (1010 CFU/ml) was collected bycentrifugation and resuspended in 0.5 ml of LB medium; XTP1was added at a multiplicity of infection of 0.1 and allowed toadsorb for 20 min at 30°C. Aliquots (10 or 100 p.l) were thendiluted in selective media (LB-Rif-Tet or LB-Rif-Kan) andgrown until lysis was complete, generally overnight.

Single-step growth experiments. The burst size and one-stepgrowth curve for XTP1 were determined as outlined bySnustad and Dean (28). XTP1 phage was added to X1231 cellsat a multiplicity of infection of 1 and allowed to adsorb for 15min at 30°C. The mixture was then diluted 10' in 5 ml of LBbroth and placed in a 30°C shaking water bath. Samples (100p.l) were removed every 15 min and titered for infective centersand unadsorbed phage (initial time point) on indicator cells(X1231) in soft agar.Adsorption of XTP1 to host bacteria. Adsorption was mea-

sured at 30°C by mixing approximately 106 PFU with 1010X1231 or BP109 cells grown in LB medium or BP109 cellsgrown in YMG. Samples were removed and assayed forunadsorbed phage by diluting 1:100 in 10% chloroform (vol/volin water), centrifuging briefly, and titering the aqueous phasefor PFU.DNA isolation and characterization. Phages were banded in

CsCl2 as described by Sambrook et al. (25). Five milliliters ofphage lysate containing 0.40 g of CsCl per ml was layered on aCsCl2 step gradient (1 ml, 0.95 g/ml; 2 ml, 0.67 g/ml; 2 ml, 0.60g/ml; 2 ml, 0.50 g/ml of CsCl2) and centrifuged overnight at

15°C in a Beckman SW41 rotor at 44,000 x g. Phage bandswere removed with a syringe and dialyzed against four changes(500 ml each) of TE buffer (10 mM Tris, 1 mM EDTA [pH 8]).To isolate phage DNA, 1 ml of phage lysate (1010 PFU/ml)

was treated with DNase (450 U) in the presence of 10 mMMgSO4 for 2 h at 37°C. The reaction was stopped by theaddition of 50 pL. of a 0.5 M EDTA stock solution. The mixturewas combined with an equal volume of 4 M guanidine hydro-chloride (G-4505; Sigma) and vortexed briefly; then a 1/10volume of 3 M sodium acetate was added. Nucleic acid wasethanol precipitated (25), washed with 70% ethanol, and dried.DNA was resuspended in 1 ml of TE-SDS buffer (25 mM Tris[pH 8.0], 0.6% sodium dodecyl sulfate [SDS], 25 p.M EDTA)by incubating at 65°C. Solubilized DNA was treated with 1 mgof fresh proteinase K at 50°C for 2 h. Nucleic acid was ethanolprecipitated, resuspended in TE (pH 7.4)-CaCl2 (10 mM)-proteinase K (2 mg/ml), and incubated at 50°C for 2 h.Proteinase K was then inactivated by heating at 70°C for 60min, and DNA was ethanol precipitated and resuspended in100 ptl of TE with RNase A (50 pLg/ml).To determine the strandedness and type of nucleic acid,

nucleic acid from the guanidine treatment (see above) wasstained with acridine orange according to Bradley (3) andtested for sensitivity to RNase-free DNase I (444 ng/pLl) andRNase A (500 ng/pl) by incubating samples for 1 h at 37°C.Digestion was monitored by agarose gel electrophoresis.DNA was tested for the presence of terminally bound

protein by treatment with exonuclease III (8 U/pl) or Xexonuclease (0.15 U/pl) in buffer (66 mM Tris, 66 mM MgCl2)for 2 h at 37°C. Digestion was monitored by agarose gelelectrophoresis.

Electron microscopy. A 3-pul drop of bacteriophage at aconcentration of 1010 PFU/ml was applied to a collodion-coated, carbon-stabilized grid. After 30 s, the drop was with-drawn by filter paper and the grid was washed with a few dropsof 1% aqueous uranyl acetate, resulting in negative staining ofthe attached phages. The coated grid was glow dischargedprior to sample application to promote better spreading of thestain. Grids were examined in a JEOL 100-CX electronmicroscope operating at 80 kV, and micrographs were re-corded at a nominal magnification of x58,000 on Kodakemulsion SO-163.

VOL. 176, 1994

on May 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

3356 WEISS ET AL.

107

106

10

0 1 2 3 4 5Time (hr)

FIG. 1. One-step growth curve for XTP1. The growth curve wasdone as described in Materials and Methods with strain X1231. Theburst size was determined by the ratio of PFU to the input bacterio-phage.

RESULTS

Isolation of bacteriophage from soil samples. Fifteen soilsamples from 12 locations in Maryland were tested for thepresence ofXanthomonas phages as described in Materials andMethods. Phages were recovered from five different locations.In each case, the phages infected both Gum+ and Gum-strains. Phages from four of the locations produced similarplaque morphologies and could transduce chromosomal mark-ers. They were assumed to be identical. One of the isolates wasdesignated XTIP1 and further characterized. Phages from thefifth location had a different plaque morphology and wereunable to transduce chromosomal markers.

Biological properties of XTP1. XTP1 forms small clearplaques of -1 mm in diameter on both Gum' and Gum-lawns. Individual plaques are more easily seen in Gum-(X1231) lawns. The addition of Ca (5 mM), Mg (5 mM), or Zn(5 mM) did not result in increased plaque size, efficiency ofplating, or phage production. Our attempts to isolate lysogenicbacteria by sampling the center of plaques were unsuccessful.XTP1 is unable to infectX campestris pv. vesicatoria (a tomatopathogen) or pv. pruni (a plum and peach pathogen) but caninfect strain X campestris pv. phaseoli (a bean pathogen).Spontaneous host range variants were not obtained in phagespottings that contained >10i phages. The plaques on Xcampestris pv. phaseoli are smaller (<0.5 mm). In contrast tothe X campestris pv. campestris host, addition of MgSO4 (5mM) improved phage yield. Attempts to transduceX campes-tris markers to X campestris pv. phaseoli XP6022 were unsuc-cessful, and transduction between X. campestris pv. phaseolistrains was not attempted.

One-step growth curves were performed as outlined inMaterials and Methods. A representative experiment is shownin Fig. 1. XTP1 has a latent period of about 2 h, which is closeto the generation time of X campestris in the exponentialphase. This is followed by a rise period of 2 h with a burst sizeof 30 to 35 PFU per adsorbed phage. The efficiency of phageadsorption can be important in transduction experiments bothin the initial stages of phage infection and in the survival oftransductants on the selective plate. Since we worked with bothGum' and Gum- strains, we assayed adsorption rates forGum- and Gum+ strains under conditions that producevarious amounts of xanthan. BP109, a Gum+ strain, producessmall amounts of xanthan in LB medium compared with

'~80.00

la 60

40

20

0

0 10 20 30

Time (min)

FIG. 2. Adsorption of XTP1 to host cells. Symbols: Ol, X1231(Gum-) grown in LB broth; A, BP109 (Gum') grown in LB broth; 0,BP109 (Gum') grown in YMG broth. Assays were performed as

described in Materials and Methods. Percent unadsorbed phage is theratio of PFU in the supernatant to the initial PFU times 100. Theinitial PFU was determined by titering an equivalent dilution of thephage in the absence of host cells.

growth in YMG medium, where xanthan gum production iselevated. Strain X1231 (Agum) does not produce xanthan gum.As shown in Fig. 2, adsorption of XTP1 to both strains isefficient. In the time necessary to plate the first samples(approximately 1 min), 50% of the phage particles adsorbed toX1231 (Gum-) and BP109/LB (low Gum) cultures. Adsorp-tion was slower in the BP109/YMG (high Gum) culture; only15% of the phages adsorbed in this same period. After 10 min,adsorption in the YMG culture was comparable to that in theother cultures (Fig. 2).

Transduction. Table 2 shows XTP1 transduction data forseveral markers of X. campestris. Transduction frequencieswere generally i0-5 to 10- transductants per PFU, and thenumber of transductants was proportional to the amount ofphage used. To be certain that colonies were true transduc-tants, purified putative transductants were shown to havegenetic markers unique to the recipient cells. We were initiallypuzzled that when recipient colonies from the Strr transduction[lysate XTP1 * BP112 (str-9 Rif') x BP109 (nif-1)] werechecked for the rif-I allele of BP109, most transductants wereRifs. One possible explanation is that the rif and str alleles arelinked in Xanthomonas species. Subsequent transduction ofboth the nif- and str-9 alleles from a Strr Rif transductant(BP125) to S4L (Rif' and Strs) showed that the alleles are atleast 75% linked inXanthomonas species (Table 2, footnote c).

TABLE 2. Generalized transduction with XTP1

Donor Relevant Avg frequency ofstraina marker transduction'

BP108 xgl-59::Tn9O3 2.5 X 10-6X1231 Agum::tet 7.5 X 10-6BP112 str-9 4.5 X 10-6BP125 str-9 (rif l)C 2.5 x 10-6BP353 pig-l::TnlO 4.5 X 10-6

a Strains are described in Table 1. Recipient strains were S4L and derivativesthat carried various genetic markers.

b Frequencies are averages of several experiments. Transduction experimentsroutinely yielded 50 to 120 transductants per plate.

c When Strr transductants were screened for Rif, it was discovered that 75%were str-9 nf-1 cotransductants.

J. BACTERIOL.

on May 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

X CAMPESTRIS PHAGE 3357

TABLE 3. Linkages of transposon insertion mutationsto xgl-4 and xgl-85

Transduction % IDonor strain' frequency of antibiotic tnxga6eitne transductantscresistance'

xgl-4 TnlO-2 0.6 X 10-6 94xgl-4 TnlO-4 1.3 x 10-6 7xgl-4 TnlO-6 1.1 X 10-6 98xgl-4 TnlO-8 7.3 X 10-6 92xgl-4 TnlO-10 0.9 x 10-6 90xgl-4 TnJO-12 2.6 X 10-6 69xgl-85 Tn9O3-14 0.9 X 10-6 68xgl-85 Tn9O3-15 1.1 x 10-6 6xgl-85 Tn9O3-16 0.7 X 10-6 49xgl-85 Tn9O3-17 1.5 X 10-6 6xgl-85 Tn9O3-18 0.8 X 10-6 6

a Donor strains were from transposon insertion pools obtained as described inMaterials and Methods with strains BP302 (xgl-4) and BP308 (xgl-85). Transduc-tants were selected on media containing Tet (for TnlO) or Kan (for Tn9O3) andX-Gal. The recipient was X77.

b Frequencies are averages of several experiments.I Ratio of pale colonies to total transductants.

To prove the utility of XTP1, we used generalized transduc-tion to genetically tag unselectable mutations with linkedselectable markers. TnlO and Tn9O3 transposon insertionpools were generated on the xgl mutants BP302 and BP308.XTP1 lysates were prepared on these pools as described inMaterials and Methods and used to transduce strain X77 toTetr or Kanr. The transductants were screened for pale yellowcolonies, i.e., those that concomitantly received the xgl muta-tion. Approximately 1 in 50 transductants received the xglmarker. Lysates were prepared on several of these xgl trans-ductants, and the transposon marker was transduced back toX77 to determine linkages to xgl. A range of linkages wasachieved with both the TnlO and Tn9O3 transposons (Table 3),suggesting that both transposons hop randomly. The linkedtransposon marker allowed us to test whether the xgl-4 allele ofBP302 was in the same locus as the xgl-59::Tn9O3 allele inBP108. Lysate prepared on BP108 was used to transduce thexgl-59::Tn903 marker to strain BP302 (xgl-4::Tn1O-10). Of the20 transductants tested, only 3 retained the Tetr determinant,demonstrating that the TnlO-10 insertion in BP302 (xgl-4) is85% linked to the Tn9O3 insertion in BP108 (xgl-59::Tn903).This is consistent with the observed linkage of 90% for theTnlO-10 insertion and the xgl-4 allele and suggests the muta-tions in BP108 and BP302 are in the same gene or gene cluster.

Physical properties. Electron microscopy of the negativelystained phages showed typical contractile tail phage morphol-ogy (Fig. 3). The phage heads are isometric icosahedral and102.4 nm in diameter. Close inspection of lysed heads revealsevidence of the head capsomers. The tail is 121.6 nm in lengthand clearly of the contractile type, showing approximately 30typical transverse tail striations. It exhibits a strikingly uniquephenomenon: at approximately one-third the length of the tailproximal to the distal (baseplate) end, there appears to be alocal thickening of the tail sheath with tail fibers emanatingradially from this site. From our observation of a number ofphages, we have not seen tail fibers originating from theconventional baseplate location at the distal end of the tail.The tail fibers appear to be of the rigid type, with a bend closeto the sheath. Images of contracted sheaths clearly reveal thenaked tail tube but, in addition, show evidence of an unfurledskirt-like sheath located below the unusual site of the tail fiberorigin (Fig. 3b). There is no evidence of a distinct baseplate

FIG. 3 Electron micrographs of the XTP1 bacteriophage. (a)Typical XTP1 bacteriophage showing the isometric icosahedral headand contractile tail with prominent tail striations. Note the uniqueorigin of the tail fibers at a location 40 nm proximal to the distal endof the tail. (b) Lysed bacteriophage showing evidence of the capsom-eric structure of the phage head and the contracted sheath revealingthe tail tube and the unfurled skirt-like region of the tail sheath belowthe tail fiber origin. (c) A typical XTP1 bacteriophage and an isolatedphage tail showing the rigid tail fibers and their origin. There is noclear evidence for a specialized baseplate structure. The proximal endof the tail shows a short neck region. Bar, 100 nm.

structure. In images of isolated tails, a short neck region is seenat the proximal end of the tail (Fig. 3c).DNA isolation and characterization. To determine the

nature and size of the XTP1 genome, we isolated nucleic acidfrom lysates. Standard approaches for isolation of nucleic acidfrom phages, including treatment with proteinase K plus SDS(25), formamide dialysis (29), sarkosyl ejection (20), precon-centration of phage particles with polyethylene glycol andCsCl2 (25), and techniques developed for X DNA (25), wereunsuccessful. We successfully isolated phage DNA throughtreatment with 2 M guanidine hydrochloride (GdnHCI) asdescribed in Materials and Methods. The nucleic acid isolatedfrom XTP1 was identified as double-stranded DNA by thefollowing criteria: (i) it stained bright green in the presence ofacridine orange according to Bradley's phage nucleic acidstaining technique (3), and (ii) it was degraded by RNase-freeDNase I but not RNase (data not shown).We observed that, when banded on CsCl2 step gradients, the

XTP1 virions yielded a dispersed band of phages that wassticky and viscous. This suggests that a polysaccharide may betightly associated with the phage. Since this was seen withlysates prepared on both Gum' and Gum- strains, the puta-tive polysaccharide is not xanthan. When lysates were grownonX campestris pv. phaseoli, the phage particles banded tightlyin the gradient, suggesting that the stickiness is due at least inpart to a host-encoded function.DNA isolated from virion particles by GdnHCI treatment

does not enter agarose gels (data not shown). Two treatmentswith proteinase K allowed XTP1 DNA to enter an agarose gel.This is indicative of a stable DNA-protein complex (24) andconsistent with the finding that phenol extraction resulted in aloss of nucleic acid from the aqueous phase unless the DNAwas pretreated with protease K. To determine if the putativebound protein(s) was associated with the 3' or 5' termini of theDNA strands, XTP1 DNA was treated with exonuclease III

VOL. 176, 1994

on May 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

3358 WEISS ET AL.

(specific for 3' ends) and X exonuclease (specific for 5' ends)(25). Both exonucleases were able to degrade purified XTP1DNA, suggesting that neither end was protected.XTP1 DNA was resistant, to digestion with restriction endo-

nucleases EcoRI, Sall, BamHL, SmaI, PstI,XhoI, and ClaI. It iscleavable by restriction endonucleases SphI, HindIII, XmnI,AatII, XbaI, Kp'nI, HpaI, EcoRV, StuI, NcoIl, and BglII. Thesize of the XTP1 genome was estimated following AatIIdigestion and separation of the fragments by agarose gelelectrophoresis. The genome is approximately 180 kb (data notshown).

DISCUSSION

We isolated a large double-stranded DNA X campestrisbacteriophage (XTP1) that is capable of generalized transduc-tion. Phages that attack different Xanthomonas strains havepreviously been characterized (34), but none capable of gen-eralized transduction .have been reported. Xanthomonasphages isolated from soil often demonstrate a broad host rangewhile those isolated from diseased plants often show a limitedhost range (12). Since we isolated similar (possibly identical)phages from four different locations in Maryland, we believethat this is a commoii phage forX campestris pv. campestris aswell as X. campestris pv.phaseoli. Why this phage has not beenpreviously identified is not known; perhaps it is geographicallylimited in its natural dispersion.Phage XTP1 has an unusual tail fiber attachment location

approximately one-third of the way up from the distal end ofthe tail. This suggests that it may have a novel mechanism ofattachment and injection of DNA. Interestingly, the contractedtail has a novel skirt-like appendage rather than the normalthickened appearance of a contracted sheath. To our kinowl-edge, such a structure has not been reported (1, 2). Its role isunknown.

Single-step growth experiments showed that XTP1 has atypical growth cycle and a burst size of 35 PFU per infectedcell. The burst size is smaller than the median of 50 to 100 fortailed phages of enteric gram-negative bacteria and pseudo-monads (1). Although adsorption was initially slower in cul-tures producing an excess of xanthan, overall adsorption ratesto Gum' and Gum- cells were similar (Fig. 2), suggesting thatxanthan does not interfere with the phage's life cycle orattachment. This contention is supported by the lack ofdifference in transduction frequencies for Gum' and Gum-recipients. This finding is in contrast to that with entericbacteria such as E. coli, in which large amounts of EPS candefend a cell from attack by bacteriophages (9). Protectionagainst bacteriophage infection may be more pronounced inbacteria that produce capsular polysaccharide, which is tech-nically distinguished from slime, i.e., unattached EPS in itsattachment to the cell surface (9).XTP1 is the first Xanthomonas phage reported to mediate

generalized transduction, and thus it can be useful for theconstruction of XA campesois strains with different geneticmarkers. We were able to show that at least five differentgenetic loci (Table 2) can be transduced between XA campestrisstrains. Accordingly, it should be possible now to move anyselectable marker into the unmutagenized background and/orgenetic background required (e.g., gum) for analysis of specificphysiological systems. In addition, this phage provides a mech-anism for the mapping of genes with respect to each other onthe chromosome. For example, we showed that str and rifalleles inX campestris are closely linked. This is different fromwhat is seen in prototypical enterics such as E. coli, in whichalleles that confer streptomycin resistance and rifampin resis-

tance are many minutes apart. Linkage of str and rif genes inthe XA campestris genome suggests that genes encoding ribo-somal proteins and RNA polymerase subunits may be adjacenton the Xanthomonas chromosome. Alternatively, differenttypes of mutations in Xanthomonas species may confer Strr andRif' phenotypes.An important XTP1 utility is that now it is possible to

genetically tag unselectable mutations with a linked antibiotic-resistant transposon (Table 3), providing a means of movingeven unselectable genetic markers. In our pool of 1,500transposon insertions, we found that about 1 in 50 transduc-tants that received a transposon also received the xgl mutation.When similar manipulations are performed with E. coli byusing generalized transducing phage P1, approximately 1 in100 transductants from a pool show linkage (31) to a specificgenetic locus. Since the XTP1 genome (180 kb) is approx-mately twice the size of the Pl genome (96 kb) and the XTP1cotransduction frequency for random transposon insertions istwice that for Pl, XA campestris and E. coli may have genomesof comparable size. Currently, we do not know how chromo-somal DNA is transmitted by the phage virion. Since we couldobtain a wide spectrum of linked transposons to the Xglmarker (Table 3 and unpublished data), it is likely that XTP1uses a headful-packaging mechanism. The inability of XTP1 totransduce genetic markers from XA campestris pv. campestris topv. phaseoli may result from restriction and modificationdifferences in the two strains. In addition, the low sequencehomology shared by XA campestris pv. campestris and XAcampestris pv. phaseoli may block interstrain transduction. Thetwo strains have only a 10% reassociation value (14).

Isolation of DNA from XTP1 was technically difficult,presumably because of two characteristics of the phage. First,XTP1 lysates were strongly associated with a sticky viscousmaterial, possibly polysaccharide. Since carbohydrates areminor components in few tailed phages (2) and only onephage, Aerobacter aerogenes phage K-2, has been shown topossess a high-molecular-weight polysaccharide (21) as anexternally exposed constituent of the phage particle, ourfindings argue that XTP1 is a novel phage. We noticed thatXTP1 lysates grown on AX campestris pv. campestris versusthose on XA campestris pv. phaseoli have different propertieswith regard to stickiness and that phage infection of pv.phaseoli was much less efficient. One explanation for thesefindings is that growth efficiency on the two strains results fromstructural variations involving a putative phage polysaccharide,the synthesis of which is dependent on host cell enzymes orsugar nucleotides. Since the phage grew on both Gum' andGum- cells, the putative polysaccharide component is notxanthan. We cannot say with certainty that the sticky materialis a component of the XTP1 capsid. The fact that it was notseparated from phage particles in CsCl gradients suggests thatit is part of the capsid rather than an associated artifact.Alternatively, the stickiness may stem from a dependency ofthe phage on EPS as its receptor. If this is the case, thereceptor is not xanthan since the phage plates efficiently on aGum- host. XA campestris pv. campestris has other EPS com-ponents in addition to xanthan (data not shown).The second characteristic that made it difficult to isolate

DNA by standard protocols was the presence of protein(s)attached to XTP1 DNA. This putative protein(s) preventedDNA from entering agarose gels and caused it to be lost fromthe aqueous layer during phenol extraction. Since treatmentsknown to denature proteins (e.g., phenol and guanidine hydro-chloride) did not break the protein-DNA interactions, webelieve the association involves covalent linkage. Several vi-ruses and bacteriophages have been reported to contain pro-

T BAcTERiOL.

on May 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

X CAMPESTRIS PHAGE 3359

teins covalently linked to the 5' end of each DNA strand (11,13, 23). However, since XTP1 DNA was degraded by bothexonuclease III and X exonuclease, it is possible that only oneend of the molecule has terminally attached protein and thatthe genome is not a closed circular molecule. Although Xexonuclease digestion of XTP1 DNA was less efficient thanexonuclease III digestion, a likely explanation is the XTP1genome's lack of protruding 5' termini, which serve as thepreferred substrate for X exonuclease (25). The nature of therole for the protein-DNA linkage remains to be determined.The large size (approximately 180 kb) of the XTP1 genome

suggests that it may be an especially useful phage for construc-tion of linked markers and genetic maps. The phage DNA maybe modified in a base- or sequence-specific manner since thephage DNA is resistant to cutting by restriction enzymes thatcut X campestris pv. campestris chromosomal DNA. Thissuggests that the phage encodes its own set of DNA modifica-tion enzymes. In comparing the recognition sequences ofrestriction enzymes that are unable to cleave XTP1 DNA, weobserved that a common theme is a 6-base sequence with5'-TCGA-3' as bases 2 through 5. It may be that the sequence5'-TCGA-3' contributes to a phage-specific modification rec-ognition sequence. However, at this time, other explanationssuch as the lack of these recognition sequences or the incor-poration of nucleotide analogs could also account for thesefindings.

ACKNOWLEDGMENTS

This work was supported by NSF grant MCB9117748 to M.C. andS.B.We thank Steven Hutcheson and Myra Jacobs for helpful comments.

REFERENCES1. Ackermann, H. W., and M. S. DuBow (ed.). 1987. Viruses of

prokaryotes, vol. I. CRC Press, Inc., Boca Raton, Fla.2. Ackermann, H. W., and M. S. DuBow (ed.). 1987. Viruses of

prokaryotes, vol. II. CRC Press, Inc., Boca Raton, Fla.3. Bradley, D. E. 1966. The fluorescent staining of bacteriophage

nucleic acids. J. Gen. Microbiol. 44:517-525.4. Capage, M. A., D. H. Doherty, M. R. Betlach, and R. W. Vander-

slice. October 1987. Patent for recombinant DNA-mediated pro-duction of xanthan gum. PCI publication no. W087/05938.

5. Civerolo, E. L. 1976. Phage sensitivity and virulence relationshipsin Xanthomonas pruni. Physiol. Plant Pathol. 9:67-75.

6. Dai, H., K. Chiang, and T. Kuo. 1980. Characterization of a newfilamentous phage Cf from Xanthomonas citri. J. Gen. Virol.46:277-289.

7. Daniels, M. J., C. E. Barber, P. C. Turner, M. K. Sawczyc, R. J. W.Byrde, and A. H. Fielding. 1984. Cloning of genes involved inpathogenicity of Xanthomonas campestris pv. campestris usingbroad host range cosmid pLAFR1. EMBO J. 3:3323-3328.

8. Ehrlich, M., K. Ehrlich, and J. A. Mayo. 1975. Unusual propertiesof the DNA from Xanthomonas oryzae phage XP-12 in which5-methyl-cytosine completely replaces cytosine. Biochim. Biophys.Acta 395:109-119.

9. Ferris, F. G., and T. J. Beveridge. 1985. Functions of bacterial cellsurface structures. BioScience 35:172-177.

10. Figurski, D. H., and D. K. Helinski. 1979. Replication of anorigin-containing derivative of plasmid RK2 dependent on aplasmid function provided in trans. Proc. Natl. Acad. Sci. USA76:1648-1652.

11. Garcia, E., A. Gomez, C. Ronda, C. Escarmis, and R. L6pez. 1983.Pneumococcal bacteriophage Cp-1 contains a protein bound tothe 5' termini of its DNA. Virology 128:92-104.

12. Goto, M., and M. P. Starr. 1972. Phage host relationships of

Xanthomonas citni compared with those of other xanthomonads.Ann. Phytopathol. Soc. Jpn. 38:226-248.

13. Grimes, S., and D. Anderson. 1989. In vitro packaging of bacterio-phage 429 restriction fragments and the role of the terminalprotein gp3. J. Mol. Biol. 209:91-100.

14. Hildebrand, D. C., N. J. Palleroni, and M. N. Schroth. 1990.Deoxyribonucleic acid relatedness of twenty-four xanthomonadstrains representing twenty-three Xanthomonas campestris patho-vars and Xanthomonas fragariae. J. Appl. Bacteriol. 68:263-269.

15. Kiew, K. W., and A. M. Alvarez. 1981. Biological and morpholog-ical characterization of Xanthomonas campestris bacteriophages.Phytopathology 71:269-273.

16. Kiew, K. W., and A. M. Alvarez. 1981. Phage typing and lysotypedistribution of Xanthomonas campestris. Phytopathology 71:274-276.

17. Liu, Y. N., J. L. Tang, B. R. Clarke, J. M. Dow, and M. J. Daniels.1990. A multipurpose broad host range cloning vector and its useto characterize an extracellular protease gene of Xanthomonascampestris pv. campestnis. Mol. Gen. Genet. 220:433-440.

18. Matthew, J., and P. N. Patel. 1979. Host specific bacteriophages ofXanthomonas vesicatoria. Phytopathology 94:3-7.

19. Miller, J. H. 1972. Experiments in molecular genetics. Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.

20. Miller, R. V., J. M. Pemberton, and K. E. Richards. 1974. F116, D3and G101: temperate bacteriophages of Pseudomonas aeruginosa.Virology 59:566-569.

21. Oeltmann, T. N., and E. C. Heath. 1975. Isolation and character-ization of a bacteriophage polysaccharide. J. Biol. Chem. 250:8696-8703.

22. Osbourne, A. E., C. E. Barber, and M. J. Daniels. 1987. Identifi-cation of plant-induced genes of the bacterial pathogen Xan-thomonas campestris pv. campestris. EMBO J. 6:23-28.

23. Rekosh, D. M. K., W. C. Russell, A. J. D. Bellet, and A. J.Robinson. 1977. Identification of a protein linked to the ends ofadenovirus DNA. Cell 11:283-295.

24. Romero, A., R. Lopez, R. Lurz, and P. Garcia. 1990. Temperatebacteriophages of Streptococcus pneumoniae that contain proteincovalently linked to the 5' ends of their DNA. J. Virol. 64:5149-5155.

25. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecularcloning: a laboratory manual, 2nd ed. Cold Spring Harbor Labo-ratory Press, Cold Spring Harbor, N.Y.

26. Sandford, P. A., and J. Baird. 1983. Industrial utilization ofpolysaccharides, p. 411-490. In G. 0. Aspinall (ed.), The polysac-charides, vol. 2. Academic Press, New York.

27. Silhavy, T. J., M. L. Berman, and L. W. Enquist (ed.). 1984.Experiments with gene fusions. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.

28. Snustad, D. P., and D. S. Dean (ed.). 1971. Genetics experimentswith bacterial viruses. W. H. Freeman & Co., San Francisco.

29. Thomas, M., and R. W. Davis. 1975. Studies on the cleavage ofbacteriophage lambda DNA with EcoRI restriction endonuclease.J. Mol. Biol. 91:315-328.

30. Thorne, L., K. Gosink, and T. Pollock. 1989. Mutants of Xan-thomonas campestris defective in secretion of extracellular en-zymes. J. Ind. Microbiol. 4:135-144.

31. Trun, N. J., and T. J. Silhavy. 1987. Characterization and in vivocloning of priC, a suppressor of signal sequence mutations inEscherichia coli K12. Genetics 116:513.

32. Vidaver, A. K., and M. L. Schuster. 1969. Characterization ofXanthomonas phaseoli bacteriophages. J. Virol. 4:300-308.

33. Vogel, H. J., and D. M. Bonner. 1956. Acetylornithinase ofEscherichia coli: partial purification and some properties. J. Biol.Chem. 218:97-106.

34. Watanabe, M., K. Naito, K. Kanebo, H. Nabasama, and D.Hosokawa. 1980. Some properties of Xanthomonas campestris pv.campestnis phage. Ann. Phytopathol. Soc. Jpn. 46:517-525.

35. Williams, P. H. 1980. Black rot: a continuing threat to worldcrucifers. Plant Dis. 64:736-742.

VOL. 176, 1994

on May 23, 2021 by guest

http://jb.asm.org/

Dow

nloaded from