Interruption-deficient mutants of bacteriophage T5 I. Isolation and general properties.

(1)

0022-538X/79/02-0716/10$02.00/0

Interruption-Deficient

Mutants of

Bacteriophage T5

I. Isolation and

General

Propertiest

STEPHEN G. ROGERS,t ELIZABETH A. GODWIN, ELIZABETH S. SHINOSKY, AND MARC RHOADES*

Departmentof Biology, The Johns Hopkins University, Baltimore, Maryland 21218

Receivedfor publication 15 August 1978

Mutants ofbacteriophage T5 were isolated which lack one or more ofthe

natural single-chain interruptions that occur in the mature DNA of this virus.

Interruption-deficient mutantswere detected by screening survivors of

hydrox-ylaminemutagenesis for altered DNA structure by electrophoresis in agarose slab

gels. Over60independentmutantswere isolated fromasurvey ofapproximately

800phageparticles.Allof themutants wereviable andcould be grouped intotwo

classes. Mutants inoneclass lackedoneof the localized siteswhere interruptions

occur in T5 DNA. To date, mutants that affect five different sites have been

obtained. Mutants inthe otherclasswereessentially free from interruptions or

hadareducedfrequency of interruptions throughoutthe genome. The members

of this class included several ambermutants. Complementation tests indicated

thatatleasttwo genes arerequired for thepresence ofinterruptions in mature

T5 DNA.

The nonpermuted genome of bacteriophage

T5 is contained within a single linear,

double-stranded DNA molecule (27). T5 DNA has a

molecular weight ofapproximately80 x 106(14,

24) and containsan 8.3% terminal redundancy

(20). T5DNA isdistinguished from the DNA of

mostothercoliphages by the presence ofasmall

number ofsingle-chain interruptions (1). These

sitescanberepairedbytheaction ofT4 DNA

ligase andarelargely confinedto onestrand of

theduplex molecule (12).Althoughthe

interrup-tions appeartooccur atfixedsites,thefrequency

ofinterruptions at the different sites is highly

variable (9, 24). The positions of the fourmost

frequent interruptions, which occur in nearly

everymolecule, have been determinedby

elec-tronmicroscopy (2, 3, 13, 23, 24) and agarose gel

electrophoresis (8, 19). These sitesoccur at

ap-proximately 8, 18, 32, and 65% from what is

arbitrarily designated asthe left end of the

ge-nome. Some of the less frequent interruptions

have also been mapped (19, 24). The highest

concentrationofthese sites occurs at the leftend

of theDNA, within the terminal repetition.

Thesignificance ofthesingle-chain

interrup-tions in T5 DNAisnotknown. Although ithas

beensuggested thatone or moreof the

interrup-tions could influence genetic recombination or

DNAinjection (1, 2),there is no directevidence

t Contribution no. 972 from the DepartmentofBiology,

TheJohnsHopkinsUniversity,Baltimore,Md.

tPresent address:Department of Microbiology, The Johns

Hopkins UniversitySchool ofMedicine,Baltimore,MD 21205.

linking these sites to any specific process. To

investigate this question, we initiated a search

formutantsof T5 that lack the normal

comple-mentofinterruptions. This searchwasaidedby

thedevelopmentofanelectrophoreticassay that

allows the genomes ofmutagenized phagetobe

rapidlyanalyzed for structural alterations. Use

of thisscreening procedure resulted in the

iso-lation of two classes of mutants with altered

patterns ofsingle-chain interruptions. Mutants

inoneclasslackedoneof the sites where

inter-ruptions normally occur. Mutants in the other

class had reduced levels of interruptions

throughoutthe genome. Some members of this

classappearedtobecompletelyfree from

inter-ruptions.

The present report describes thegeneral

prop-erties of interruption-deficient mutants of T5

and the methods used for their isolation. The

accompanying paper (21) describes the

proper-ties of a mutantthatspecificallylacks the

inter-ruption located at 8% from the left end of T5

DNA.

MATERIALS AND METHODS Bacteriaandbacteriophage. T5st(+)isthe wild-typestrain of Abelson and Thomas(1). T5st(102) isa

heat-stable deletion mutant (25). Escherichia coli B40sul, B40su2, andB40su3 wereobtained from H. Berger.E. coliF,afast-adsorbing strainforT5,was

obtained from D. Botstein. E.coli B23 wasobtained from M. Bessman. All bacteriophage weregrownby confluentlysisasdescribedbyRhoadesandRhoades (20).

716

on November 10, 2019 by guest

http://jvi.asm.org/

(2)

Isolationofmutants.T5st(102)wastreatedwith hydroxylaminefor72h at370Casdescribed by Tess-man (26) to afinalsurvival of106. During the same period, a control incubated without hydroxylamine showed a 50% loss in titer. Survivors were plated on E. coliB40sul at 30°C. Individual plaques were sus-pended in 10 mM Tris-hydrochloride (pH 8.0) and replated on E.coli B40sul to eliminate heterozygotes. Single plaques were then transferred with sterile toothpicks to 1-cm2 grids on soft-agar overlays (6) seeded with E. coliB40sul. After 6to 8 h at 30 or

370C,thesoft-agar grids were removed from theplates

with aspatula andplaced in 0.1 ml of 10 mM Tris-hydrochloride (pH 8.0). After 30 min at250C,a 0.03-mlsample of the resuspended phage was mixed with 5,Aof 2% sodiumdodecylsulfate-0.1 M EDTA (pH 9.0) and heatedfor 5 min at 60°C. Ten minutes before analysis of the DNA, 5

pl

of afreshly prepared solution of1 MNaOH was added to denature the DNA. The denaturedsamples were then subjected to electropho-resis inhorizontal 0.7% agarose slab gels as described by Rogers and Rhoades (22).

Amber mutants wereobtained byplating survivors ofhydroxylaminetreatmentonthe su- strain E. coli B23. Single plaqueswerethen directly tested for in-terruption-deficient phenotypes as described above. Phage which were either completely free from inter-ruptionsorwhichexhibited reducedlevels of interrup-tionsthroughout the genome were then retested on E. coliB40sul, B40su2, and B40su3.

Procedurefor crosses. E. coli B40sul or E. coli Fwasgrown to 2x108 cells/mlinT-MGM (15) with aeration at370C. Aftercentrifugation, the cells were resuspendedat109/mlin TMB (T-MGM without glu-cose or NH4C1) containing 1 mM CaC12 and were starved for 20min at 370C with aeration. The cells were then diluted with an equal volume of TMB containing1 mMCaC12andwereinfected at a multi-plicity of 7.5 of each parent. After 20 min at 37°C without aeration, the infected cells were centrifuged to remove unadsorbedphage andweregently resus-pendedat 5x 108/mlinprewarmed T-MGM contain-ing 1 mM CaCl2.After 5 min at370C, aeration was started and continued for 60 to 90 min. Lysis was completed by addition ofCHC13.Progeny wereplated on an appropriate indicator, and the nature of the DNA of individual plaques was tested as described above.

In crosses where one parent carried the st(102) deletion mutation, the progeny were fractionated either by plating on E. coli F, where the st(102) mutation confers a small-plaque phenotype, or by banding in CsCldensity gradients. In the latter case, aportion of thelysate wasmade 45% (wt/wt) CsCl andcentrifugedin aBeckmanSW50.1rotor at25,000 rpm for16 h at40C.

Complementation analysis and dominance tests.Lysates ofmixedly infected bacteria were pre-paredasdescribedabove, except thatE.coli B23was used for all experiments involving amber mutants. Electrophoresiswascarried outdirectly on the prog-enypresentin thelysates.In somecases,particularly if oneparent carried thest(102) deletion, the progeny werefirstfractionatedby centrifugation in CsCl den-sity gradients. After dialysis against 10mM

Tris-hy-drochloride (pH8.0)-i mMMgCl2-1 mM CaC12, the CsCl-purified phage particleswere analyzed by elec-trophoresis.

RESULTS

Isolation of interruption-deficient mu-tants. The presence of single-chain

interrup-tions in T5 DNA is not known to confer any

phenotypicpropertiesuponthe virus.

Interrup-tion-deficient mutants were therefore obtained

bydirect screening ofmutagenized phage

prep-arations. To survey rapidly a large number of

potential mutants, we used a procedure that

permits theDNA presentinasingle T5 plaque

-to be analyzed, after alkaline denaturation, by

agarosegel electrophoresis. Since bacterial DNA

isefficiently degraded during the earlystagesof

T5infection (5),virtually all of the DNA inaT5

plaque consists ofphage DNA, either packaged

orunpackaged.The DNAcontentofawild-type

T5 plaque issufficientlyhigh (0.01to0.1 jug) to

permitelectrophoretic analysiswith the use of

ethidiumbromidestaining.Inpractice,however,

more satisfactory results are obtained if each

plaque isreplicaplatedtoform a 1-cm2area of

lysisasdescribed in Materials and Methods.

Mutagenesis was carried out by treating

T5st(102) phage particles with hydroxylamine

until thesurviving fractionwas 10-6. T5st(102),

aheat-stable deletionmutantthatlacks 10% of

thewild-typegenome(25), waschosento

mini-mizeanylossof viablephagenotassociated with

the action of hydroxylamine. The surviving

phage particles were analyzed by two

proce-dures. In theinitialstagesof thisinvestigation,

survivorswereplatedat

300C

onE. coli B40sul

to avoid loss of conditionally lethal mutants.

Heterozygotes wereeliminatedbyreplating

in-dividualplaquesunder identical conditions.

Sev-eral plaques derived from each originalisolate

werethenanalyzedforloss ofinterruptions.

Theresults ofa typical analysisof22

prepa-rations, derived from 10 independentsurvivors

ofhydroxylaminemutagenesis,areshowninFig.

1.With theexceptionof slot23, which received

aninsufficientamount ofDNA, all of the

sam-plesexhibiteda distinctpattern ofsingle-chain

fragments.A mutantpatternwasclearlyevident

in slots 7and 8. Both of these samples, which

were derived from the same isolate, HA99,

lacked two of theprominent single-chain

frag-ments that were present in all of the other

samplesand inpurifiedT5st(102)DNA(slot 1).

Asdescribedbelow, thisparticularmutant

pat-tern indicates lossof the interruption at18.5%.

Ofapproximately 150survivorsof hydroxyla-mine treatment that were screened in this man-ner, 16 were found to contain atleastone mu-tation which caused an interruption-deficient 29,1979

http://jvi.asm.org/

(3)

5 10 15

T v _ _

N

-

.

t..H.... s~~

~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~.Ai

~

.i

a.

A

1ll s 1 I I I |~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~F 1

FIG. 1. Exampleof isolation of interruption-deficient T5 mutants. The selection shown contained two or three descendants of isolates HA92through HA101 from hydroxylamine mutagenesis of T5st(102). Both descendants ofHA99 showed a mutant single-chainpattern in slots 7and 8. Slot 1 contained purified T5st(1 02) DNA. The other22sampleswereprepared for electrophoresis from small-platelysates as described in Materials and Methods. The directionof electrophoresis in this and all other figures is from top to bottom.

phenotype. In one instance, HA104, two

differ-entmutantphenotypessegregatedupon

replat-ing. Three different mutant phenotypes were

recovered from a second isolate, HA73, by

re-combination (see Table 1).

Interruption-deficientmutantsof T5werealso

obtained by a more rapid procedure in which

survivors of the hydroxylamine treatment

de-scribed above weretested directly, without

re-plating,for loss ofinterruptions.E. coliB23, an

su host,wasemployedin theseexperimentsin

order to allow isolation of amber

interruption-deficient mutants. Approximately 650 plaques

werescreened in thismanner, with the isolation of 45 additional mutants. The phenotypes of four of these mutants were found to be sup-pressed by growth on su+ strains.

Members of each of the categories of mutants obtained (see Table 1) were crossed with the wild type in an attempt to eliminate extraneous

mutations. Theoriginalisolates,whichcarry the

st(102) deletion,werecrossed withT5st(+),and

interruption-deficient st(+) recombinants were

obtained. The interruption-deficient mutation

was then crossed back and forth between

T5st(+) andT5st(102)for,in some cases, atotal

of five backcrosses. Wheneverpossible,any

de-tectable conditionallethal, plaque morphology,

orhostrangemutationswereeliminatedby this

procedure.However,as aresult of thehigh level

ofmutagenesisemployedinthisstudy, it islikely

that none of the strains are isogenic with the

wild type.

Electrophoreticanalysis.Thevarious

mu-tantpatternsofsingle-chain fragmentsthatwere

obtainedareshown inFig.2and3.For

conven-ience, all of the phage shown carrythe st(102)

deletion. A map showing the positions of the

single-chain interruptions and fragments that

are relevant for this study is shown in Fig. 4.

The fragments are numbered as described by

Rhoades(19).

The fragment patterns of the five mutants

shown inFig.2all differedfromT5st(102)DNA in onlya few instances. The simplest

interpre-tation of each of these patterns is that they

http://jvi.asm.org/

[image:3.505.68.458.66.381.2]

(4)

2 3

rw

0

t _4

* * * _~~~"

_ _$

4 5 6 7 is probably fragment 4 inT5HA65 (slot 4) sug-gests loss of the interruption at 32.6%. A new fragment that presumably representsthe fusion

of fragments 4 and 13' is visible just behind

fragment3.Theinterpretationof T5HA214 (slot

6) is lessclear; however,thediminished intensity

of the band where fragments 3 and 4 nonnally comigrate and the absence offragment 15 sug-gestthatthe

interruption

at64.8% has been lost.

The faint band remaining at the position of

fragments 3 and 4 probably represents a

frag-3 4 mentthatextends from 45%to an interruption

1 2 3 4

e!w

---nr.-

ew 8,9

15 16

18

21

[image:4.505.56.246.72.463.2]

13

FIG. 2. Electrophoreticpatternsof five single-site interruption-deficientmutants.All slots contained1

pg of alkali-denatured DNA from CsCl-purified phage particles.Slots 1 and7, T5st(1 02);slot2,HA4; slot3, HA626;slot4, HA65;slot5, HA73-16; slot6, HA214. The locationsofthe numberedfragmentsare

shown inFig.4. Thesymbol"i"designatesthe intact

strandofT5 DNA.

reflect loss ofoneof the siteswhereinterruptions normally occur in T5 DNA. In the case of T5HA4 (slot 2),the absence offragment 18 and

some of the smaller fragmentssuggests loss of

theinterruptionat7.9%.Moreover,severalnew

fragments which span the site of the missing

7.9% interruption arevisible (21). The absence

offragments13' and 18 in T5HA626(slot 3),and the increased intensity of fragment 7', both in-dicate loss ofthe interruption at 18.5%. Simi-larly, the absence of fragments 13', 16, and what

&-.:

I 4

18

21

13

FIG. 3. Electrophoreticpatterns of four mutants that exhibited reduced levels of interruptions throughoutthegenome.Allslots contained0.53 gof alkali-denatured DNAfromCsCl-purifiedphage

par-ticles. Slots 1 and6, T5st(102); slot2, HA23; slot3, HA104S;slot4,HA90;slot5, HA229.

5 6

29,1979

I

http://jvi.asm.org/

[image:4.505.270.436.210.609.2]

(5)

0.8 7.9 18.5 32 6

21 8 13_

I 1--

-1-22 7

I l

64.8 77.9

45.3

4

[image:5.505.111.413.69.148.2]

16 9

FIG. 4. Maps of single-chaininterruptionsandsingle-chainfragments ofT5 DNA. The locationsofseven interruptions, expressedas apercentageof thewild-type length from the left end of themolecule,aregivenon the top line. The locationsof themoreprominent fragmentsof T5 DNA that define these sitesaregiven in the lower partofthefigure. Onlythe sites andfragmentsthat arerelevantforthisstudyare shown. A more detailed map is available in the accompanying paper (21). The st(102) deletion extendsfrom21.6to32.0%o. Fragments7' and 13'areshortenedversionsoffragments7and13thatoccurinT5st(+) DNA.

thatoccurswithlowfrequencyat77%(19).The

newfragment formedby the fusion of fragments

3 and 4 would not be resolved fromthe intact

strand under these conditions. Thesefour

mu-tantsthusrepresent,individually, loss of each of

the four principal single-chain interruptions in

T5 DNA.

The fifthmutantshown inFig. 2,T5HA73-16

(slot 5), was isolated by recombination from

T5HA73 (see Table 1). Although HA73-16 has

the fourprincipal interruptions, the absenceof

fragment 16 suggests loss of the interruption

thatnormallyoccurs at45%. Loss offragment9,

which would also be expected inHA73-16, was

difflculttodetect sincefragments8and9

comi-grateunder these conditions.

Additional electrophoretic studies, involving

analogous sets of duplex fragments (19), were

performed onthe fivemutants shown in Fig. 2.

In eachcase,theinterpretationgiven abovewas

confirmed. The results of this kind ofanalysis

with T5HA4 are given in the accompanying

paper (21).

The mutant patterns shown in Fig. 3reflect

reduced levels of interruptions throughout the

genome. In the case of T5HA23 (slot 2), the

absence of detectable fragments smaller than

the intact strand indicates that this genome is

virtually free from interruptions. T5HA104S

(slot 3) appears to retain some interruptions

although the overall level is greatly reduced.

The patternsof T5HA90 and T5HA229 (slots4 and 5) both suggestthat mostofthe principal

interruptions are present, but thatthe less

fre-quent secondary interruptions are either

com-pletely absent (HA90) or greatly reduced

(HA229). The most prominent feature of

T5HA229 is the increasedintensityoffragment

21.The newbands that canbeseeninT5HA90

andT5HA104Sareprobably derived from

mol-ecules that lack either orboth of the principal

interruptionsat7.9and 18.5%.

The patterns ofsuppressionof thefouramber

interruption-deficientmutantsareshown inFig.

5. Two of these mutants, T5amHA676 and

1 2 3 4 S 6 7 8 9 10 11 12

LIlUIX 4i.

FIG. 5. Electrophoretic analysis of amber inter-ruption-deficient mutants after growth on su- and su+strains. Slots1 to 4containedT5amHA676 grown on E.coliB23, B40sul, B40su2, and B40su3, respec-tively. Slots5 to8containedT5amHA911 grown on thesamehosts.Slots9 and10 containedT5amHA346 grown onB23 and B40sul. Slots 11 and 12 contained T5amHA910 grown on B23 and B40sul. All slots contained alkali-denatured DNAfromCsCl-purified phage particles.

T5amHA911,werefree frominterruptionswhen

grown on E. coli B23. After growth onE. coli

B40sul andB40su3, amHA676wassuppressed

tothe wildtype(slots2and4); however,growth

J. VIROL.

http://jvi.asm.org/

[image:5.505.274.465.217.522.2]

(6)

on B40su2 resulted in a pattern similar to that ofT5HA90 (see Fig. 3) in which only the prin-cipal interruptions are present (slot 3). In con-trast,amHA911 was fully suppressed only by E. coli B40su2 (slot 7). Growth on B40sul resulted in a reduced level of interruptions (similar to T5HA104S), whereas no suppression occurred on B40su3. The two other amber mutants, T5amHA346 and T5amHA910, exhibited a re-duced level ofinterruptions on E. coli B23 (slots 9and11) that resembles HA104S. Both of these mutants weresuppressed tothe wildtype onE. coli B40sul(slots10and12) and on the su2 and su3 strains (datanotshown). In addition, all of

the mutants that had a phenotype on E. coli

B40sulthat was intermediate between the wild

type and completely interruption deficient (like HA90 andHA104S) were tested on E. coli B23. None ofthepatterns werefoundtochange.

A list of the categories of

interruption-defi-cient T5 mutantsobtained in thisstudyisgiven in Table 1.

Dominancetests.Dominancetestswere

per-TABLE 1. Categoriesof interruption-deficient mutants

Mutantphenotype' Nlateidb Representativeallelesc

8- 7(2) HA4,HA73-17, HA336

18- 11 (1) HA99, HA626, HA682

32- 16(4) HA65, HA77, HA435

45- 2 (1) HA73-16, HA772

65- 5(3) HA83, HA214, HA392

Interruption-free 9(4) HA23, HA34, HA54, HA104L, HA363, HA663, amHA676, HA883, amHA911 Reducedinterrup- 9(2) HA73, HA104S,

tions amHA346, HA645,

HA771, HA776,

HA896, amHA910, HA935

Principal inter- 3(1) HA90, HA599, HA835 ruptions only

Reduced second- 1 (1) HA229 ary

interrup-tions

aThedesignations "8-," "18-," etc., refer to phage specifically lacking interruptionsat8%, 18%, etc., from the left end of T5 DNA. The last four categories describe mutantswithgenerallyreduced levels of in-terruptions throughoutthegenome,as shown inFig. 3.

b The values inparenthesesarethenumber of mu-tantsobtained inthe initialphase ofthisstudy when each isolate was replated before testing. All allele numbers below HA300aremembersof thisgroup.

cMutantsHA73-17andHA73-16wereisolatedfrom HA73 by recombination. HA104L and HA104S are segregants of isolate HA104. Ambermutants are de-notedby theprefix"am."

formed by analyzing the single-chain fragments present inthe unreplicated progeny of mixedly infected bacteria. In thecaseof themutantsthat lackedasingle interruption, mixed infection with the wild type gave bursts thatappeared to con-tain both mutant andwild-type genomes. The results obtained for mixed infection with T5st(+)HA4, which lacks the interruption at 7.9%, and T5st(102) are shown in Fig. 6. In

1 2 3 4

131

[image:6.505.277.422.182.586.2]

18 ? - _ _ _ _ _

FIG. 6. Electrophoreticanalysisoftheprogenyof mixedinfectionwithT5st(+)andT5st(102)HA4. The progeny werefractionated in CsCl to separate the st(+) and st(102) phage before analysis. Slot 1, T5st(+) DNA; slot2, T5st(102) DNA; slot 3, mixed infectionprogenycarryingthest(102)deletion;slot4, T5st(102)HA4DNA. Thesymbols"a" and "b" denote two newfragments that appeared in T5st(102)HA4 DNA.

VOL. 29,1979

http://jvi.asm.org/

(7)

722

addition to lacking fragment 18, T5HA4

con-tainedseveralnewfragmentsthatdistinguish it

from thewildtype (Fig. 6,slot4). As described

in the accompanying paper (21), the new frag-ments extended between interruptions located

closetotheleft end ofgenomeandthe

interrup-tionat18.5%. Thepresence of bothmutantand

wild-typegenomes wastherefore demonstrated

by the presence of both fragment 18 and the

newfragments in the mixed infection progeny.

Similar results were obtained for T5HA99,

which lacks the interruption at 18.5%, and for

T5HA65,whichlacks the interruption at32.6%.

Analysisof themutantsthat lack interruptions

at45.3and64.8%wasmoredifficult since, under

the electrophoretic conditions employed here,

neither of these mutants possesses fragments

thatareabsent inthe wildtype.

These results indicate that loss of a single

interruption isnotduetothe absence ofa

dif-fusiblegeneproduct.Instead, itappears thatin

eachmutant oneofthe sites where interruptions

normallyoccur has beenalteredsothatthesite

isno longerrecognized by the mechanism that

specifies the interruptions.

In contrast,all of the mutations that resultin

areducedlevel ofinterruptions throughout the

genome appear to be recessive to wild type.

These tests werecarriedout onthe 21mutants

listed in Table 1 thatresemble eitherT5HA23

(no interruptions) or T5HA104S and T5HA90

(reducedlevels ofinterruptions).Inallcases,the

progenyof mixedly infected bacteriawerefound

topossessnormalpatterns ofsingle-chain

frag-ments. The results of mixed infection with

T5st(+)HA23 and T5st(102)areshownin Fig. 7.

T5HA229alsoappears tobe recessive;however,

the phenotype of this mutant is so similar to

that ofthe wild type (see Fig. 3) that an

inter-mediate result would be difficult to recognize.

Complementation analysis. The results

presented above suggest that absence ofall or

most of the single-chain interruptions in T5

DNA canbeduetothelack ofadiffusiblegene

product. To determine the number ofgenes

in-volved in this process, we carried out comple-mentation tests between various pairs of

inter-ruption-free andreduced-interruption mutants.

Theresultsof theseexperiments,which are

sum-marized inTable2, can be stated asfollows.In

all of the tests involving pairs of

interruption-free mutants, no complementation was ob-served.The progeny, like the parents, were free

frominterruptions. Severalofthe mutants that

had reduced levels of interruptions (HA346, HA645, HA771, amHA910, and HA935) also failed tocomplementtheinterruption-free mu-tants.In thesecases, theprogeny resembled the

1 2

I..

A

FIG. 7. Electrophoretic analysis oftheprogenyof

mixedinfectionwithT5st(+)HA23andT5st(102).The progeny werefractionated in CsCl to separate the st(+) andst(102)phage before analysis.Slot1, mixed-infectionprogeny of st(102) density; slot 2, mixed-infectionprogenyof st(+)density.

reduced-interruption parent.However, threeof

the reduced-interruption mutants (HA73,

HA104S, and HA896) were found to partially

complement the interruption-freemutants and

theotherreduced-interruptionmutants.The re-sults of mixed infections with T5HA104S and

several of the interruption-free mutants are

shown inFig.8.In eachcase, themixed-infection

progeny hadall of the major single-chain

frag-ments, but the minor fragments were barely

visible. These twogroups of mutants can,

there-fore, cooperate to produce progeny that have

mostbut notallof theinterruptions present in the wild type. Oneinterpretationof this result is

http://jvi.asm.org/

[image:7.505.339.387.76.462.2]

(8)

TABLE 2. Complementation analysis ofinterruption-deficientmutantsa

Allele Allele

HA23 HA34b HA363' amHA676 HA73 HA104S amHA346c HA896

HA896 +

amHA346c - +

HA104S + + + + -

-HA73 +

amHA676

-HA363C

-HA34b - _

HA23

-a+, Positivecomplementationas

shown

inFig. 8; -,nocomplementation.

bResults identical to that shown for HA34 have been obtained with HA54 and HA104L.

All

pairwisetests within this group werenegative.

cResults identicalto that shown for amHA346 have been obtained with HA645, HA771, amHA910, and HA935. Results identical to that shownfor HA363 have been obtained for HA663, HA883,and amHA911. Pairwise tests within these two groups have not been done.

that T5HA23 and T5HA104S define two genes whose functions are required for wild-type levels ofinterruptions. We have tentatively designated

the gene represented by HA23 as sciA

(single-chaininterruption) and thegenerepresentedby HA104S as sciB.

Otherproperties. Thegrowth properties of

T5HA23, which is free from interruptions, were

tested for any detectable differences from the

wild type. In addition to normal plaque

mor-phology, the latent period and burst size of

T5HA23 were indistinguishable from the wild

type.The absence ofinterruptions also did not

alter the survival of either T5st(+)HA23 or

T5st(102)HA23 after heatingto60°C incitrate

buffer(11).

DISCUSSION

This report describes the isolation of two

classesof viable interruption-deficient mutants

ofbacteriophage T5.In one class,the

interrup-tion deficiency appears to be limited, in each

mutant, to a single site. Since these mutations

are neither dominantnor recessive tothe wild

type, theypresumably represent alterations of

criticalsequences thatspecifytheinterruptions.

Mutants in the second class are

interruption-free or have reduced levels of interruptions

throughout the genome. These mutations are

recessivetowildtypeandappeartodefinetwo

genes, sciA andsciB,whoseproductsparticipate

informationoftheinterruptionsin T5 DNA.

Bothclassesofinterruption-deficientmutants

wereisolatedfrom a stock ofT5st(102)that had

beenheavily mutagenizedwithhydroxylamine.

Althoughthe resulting highfrequencyof

inter-ruption-deficientmutants wasadvantageousfor

this study, it is very unlikely that any of the

surviving phage were free from additional

hy-droxylamine-induced mutations. We have

at-temptedto reduce the levelof extraneous

mu-tations by backcrossing selected mutants with

the wildtype.Thest(+) and st(102) alleleswere

employedassecond markersinthese crosses to

insure that the mutant genomes were able to

recombine. Since this procedure may neverbe

completely successful, it is fortunate that

mul-tiple alleleswereobtainedformostof themutant

categories. Future studies of each

interruption-deficient phenotype will therefore not be

de-pendent upon the possibly aberrant properties

ofsinglemutants.

Theoccurrenceof both classes of

interruption-deficientmutants atnearly equal frequenciesis

of interest since thesingle-sitemutants

presum-ablyrepresentalterations ofsequencesthatare

much smaller thanagene.Hydroxylamine,

how-ever, is known to mutate preferentially

single-stranded DNA (7).Since the regions

surround-ing interruptions are more likely to be

tran-siently single stranded than other regions

(es-peciallyinaphage particle), hydroxylaminemay preferentiallymutatethese sites in T5 DNA. If correct, thishypothesis offersa possible

expla-nation for the observationthat,with one

appar-entexception (M. Rhoades, unpublished data),

no mutants were obtained with new

interrup-tions.

Informationonthesequencessurrounding the

interruptions in T5DNAhasrecentlybeen

ob-tainedbyNichols and Donelson(17,18). Astudy

of the 5' sequences, done on intactDNA, indi-catesthatpGpCpGpC-occurs atallof the

inter-ruptions and suggests that the octanucleotide

pGpCpGpCpGpGpTpG- occurs at most orallof

theprincipal interruptions.Thesequencesatthe

3' side are more diverse. Studies on isolated

single-chain fragmentshaveprovidedextensive

29,

http://jvi.asm.org/

(9)

724 AL.

1 2 3 4 5

vL av~

-6 7

-'.-i

Pvapt

3,4

zl

;18

13

8

[image:9.505.61.251.64.412.2]

VW.';

FIG. 8. Electrophoretic analysis of complementa-tiontestsbetweenT5HA104S(reduced interruptions) and several interruption-freemutants. Slot1, dena-tured T5st(102)HA54 DNA; slot 2, denatured T5st(102)HA23DNA;slot3,denaturedT5st(+)HA23 DNA;slot4,denaturedT5st(102)HA104S DNA;slot 5, denatured T5st(102) DNA; slot 6,progeny from mixed infection with T5st(102)HA104S and T5st(102)HA54; slot 7, st(102) densityprogeny from mixed infection with T5st(102)HA104S and

T5st(+)HA34; slot 8, st(102) densityprogeny from mixed infection with T5st(102)HA104S and T5st(+)HA23.

3' sequences at two sites, 8 and 18%, where interruption-deficient mutants have been ob-tained. At the 18%sitefourofthe first six base pairsonthe 3' sideareGC,whereasat8%oneof

thefirst six isaGCpair. Therearethus numer-oussites inthe immediatevicinity of thesetwo interruptions where hydroxylamine, which is specific for C, could act. Thehigher frequency ofGC pairsat 18%, however, doesnotseemto have significantly increased the frequency of mutations atthis siteascomparedto8% (Table 1). It is possible, as suggested by Nichols and

Donelson (18), that the critical 3' sequences

occurfarther fromtheinterruptions inaregion

whereoverlapping segmentswithtwofold

rota-tionalsymmetry arefound.

The existence ofviable T5mutants that are

apparently free frominterruptions raises

ques-tionsastowhetherthese sites playanessential

role in the viral life cycle. The DNA of these

mutants,however, hasonly been showntolack

interruptions inmaturevirus particles. The

in-tracellular forms of the DNAof the

interruption-deficientmutantshavenotbeenexamined. It is

possible that thedefects seenin these mutants

affect only the terminalstagesof viral

develop-ment. Therefore, atpresent, theonly firm

con-clusions are that intact T5 phage heads can

contain DNA that hasnointerruptions and that

such DNA is infective.

Theconclusion thatatleasttwogenescontrol

the formation of interruptions in mature T5

DNAwasbasedonthepartialcomplementation

observed after mixed infection with certain pairs ofmutants.Sinceafullwild-type phenotypewas never observed, these results might be due to

intragenic complementation. However, all

com-plementing pairs invariably produced patterns

similartothose shown inFig.3,including

infec-tionson su- hosts in which oneparent carried

anamber mutation. Thus, althoughthereason

for partial complementation is notknown, the

presentresults favoratwo-genehypothesis.

Theavailability ofinterruption-deficient

mu-tantsof T5should greatly facilitate further

anal-ysesof the origin and function of the

interrup-tions in T5 DNA. In particular, the single-site

mutantsshouldpermitevaluation of the role of

interruptions inprocessessuchasgenetic

recom-bination, transcription control, and DNA

trans-fer. As described in the accompanying paper

(21), T5HA4, which lacks the interruption at

8%, isnotdeficient in thetwo-steptransfer

proc-ess that is uniquetoT5. Itshould also be

pos-sibletorecombine the varioussingle-site

muta-tions to produce multiple mutants. Since the

defects in the single-site mutants appearto be

permanent, the construction of recombinants that lackallof theprincipal interruptions should

provide additional insights into the possible

functions of these sites.

Analysis of the interruption-free mutants

shouldprovidesomeinsightinto the mechanism

that generatesinterruptionsin T5 DNA. Little

is known about the nature of the T5 sciA and

sciB genes, although bacteria infected with an sciA mutant, HA23, are not deficient for four

previously isolated (22)site-specific

endonucle-asesthat are inducedbyT5 (S.Rogers,

unpub-lished data). The availability of amber

muta-tions in thesciA gene shouldpermit identifica-tion and isolaidentifica-tion of its geneproduct.Knowledge J. VIROL.

http://jvi.asm.org/

(10)

of thelocation ofthescigenes onthe T5genetic map may also provide some indications as to

how, or when, theyareexpressed. Preliminary

results (B.Lange-Gustafson, unpublisheddata)

havelocated the HA23 mutationin alategene

region atthe extremeright of the nonrepeated

portion of the T5genome.None of the mutations

in the sciB gene has been mapped; however,

recombinationbetween HA23 andHA104S

ap-pears tobeinfrequent.

The T5 D15gene, anessentialearlygenethat

specifies a 5' exonuclease (10), has been

impli-cated ingeneratinginterruptions in intracellular

T5 DNA. Although D15 mutants are defective

ininitiatinglategeneexpression (4), Moyer and

Rothe(16) have obtained evidence that the D15

gene product directly participates in forming

interruptions. These authors have also

discov-ered that thepurifiedD15 proteinhas

endonu-cleolyticactivity, although it is unableto

intro-duce single-chain interruptions into duplex

DNA. Itis alsonotknown whether the breaks

generated by the D15 nucleasearethosepresent

in mature phage particles. Nevertheless, these

resultssuggestthat theprocess orprocessesthat

place interruptions into T5 DNA are complex

and that theirunderstanding willrequire

anal-ysisofanumber ofgenes.

ACKNOWLEDGMENTS

Thisresearchwassupported bygrantPCM76-24036from theNational Science Foundation. S.G.R.wassupported bya training grant (HD-139) from the National Institutes of

Health.

LITERATURE CMD

1.Abelson,J.,and C. A.Thomas,Jr.1966.Theanatomy of theT5bacteriophage DNAmolecule. J.Mol. Biol. 18:262-291.

2. Bujard,H. 1969. Location ofsingle-strand interruptions

inthe DNA ofbacteriophageT5+. Proc. Natl. Acad. Sci. U.S.A. 62:1167-1174.

3. Bujard,H., and H. E. Hendrickson. 1973. Structure

and function of the genome ofcoliphageT5. I. The

physicalstructureof the chromosome ofT5+.Eur. J. Biochem. 33:517-528.

4.Chinnadurai,G.,andD. J.McCorquodale. 1973.

Re-quirement ofa phage-induced 5'-exonuclease for the

expressionof late genesofbacteriophageT5. Proc.Natl. Acad.Sci.U.S.A.70:3502-3505.

5. Crawford,L. V.1959.Nucleic acid metabolism in Esch-erichiacoli infected withphageT5.Virology7:359-374. 6. Doermann,A.H.,andL Boehner.1970.The identifi-cation ofcomplex genotypes inbacteriophage T4. I. Methods. Genetics66:417-428.

7.Freese, E.,and H. B.Strack.1962.Induction of muta-tions intransforming DNAbyhydroxylamine. Proc. Natl. Acad.Sci. U.S.A.48:1796-1803.

8. Hayward,G.S.1974.Unique double-stranded fragments ofbacteriophage T5 DNA resulting frompreferential shear-induced breakage at nicks. Proc. Natl. Acad. Sci. U.S.A. 71:2108-2112.

9. Hayward, G. S., and M. G. Smith. 1972. The chromo-some ofbacteriophage T5. I. Analysis of the single-stranded DNAfragments by agarose gel electrophore-sis. J.Mol.Biol. 63:383-395.

10.Hendrickson, H.E., and D. J.McCorquodale. 1972. Geneticandphysiological studies of bacteriophage T5. III.Patternsofdeoxyribonucleic acid synthesis induced by mutants of T5 and the identification of genes influ-encing theappearance of phage-induced dihydrofolate reductase and deoxyribonuclease. J. Virol. 9:981-989. 11. Hertel, R., L. Marchi, and K. Muller. 1962. Density

mutants ofphage T5.Virology 18:576-581.

12. Jacquemin-Sablon, A., and C. C. Richardson. 1970. Analysis of the interruptions in bacteriophage T5 DNA. J.Mol.Biol. 47:477-493.

13.Johnston, J.V.,B. P.Nichols, and J. E. Donelson. 1977.Distributionof "minor" nicks in bacteriophage T5 DNA. J. Virol. 22:510-519.

14. Lang,D., A. R.Shaw,and D. J. McCorquodale. 1976. Molecular weights of DNA from bacteriophages T5,

T5st(0),BF23st(4). J.Virol. 17:296-297.

15.Moyer, R. W., A. S. Fu, and C. Szabo. 1972. Regulation ofbacteriophage T5 development by Col I factors. J. Virol. 9:804-812.

16. Moyer, R.W., andC.T. Rothe. 1977. Role of the T5 geneD15 nuclease in the generation of nicked bacterio-phage T5 DNA. J. Virol. 24:177-193.

17. Nichols, B. P., and J. E. Donelson. 1977. Sequence analysis of the nicks and termini of bacteriophage T5 DNA. J.Virol. 22:520-526.

18. Nichols,B.P., andJ. E. Donelson. 1977. The nucleotide sequence at the 3'-termini of three major T5 DNA

fragments.Virology83:396-403.

19.Rhoades,M. 1977.Localization of single-chain interrup-tions in bacteriophage T5 DNA. II. Electrophoretic studies. J. Virol. 23:737-750.

20.Rhoades, M.,and E. A. Rhoades. 1972.Terminal repe-tition in the DNA of bacteriophageT5.J. Mol.Biol.

69:187-200.

21. Rogers,S.G.,N. V.Hamlett, and M. Rhoades. 1979. Interruption-deficient mutants of bacteriophage T5. II.

Propertiesof a mutantlacking a specific interruption. J.Virol. 29:726-734.

22. Rogers,S.G., and M. Rhoades. 1976. Bacteriophage T5-induced endonucleases that introduce site-specific

single-chain interruptionsinduplex DNA. Proc. Natl. Acad. Sci. U.S.A. 73:1576-1580.

23. Saigo,K.1975.Denaturationmapping and chromosome structureinbacteriophageT5.Virology68:166-172.

24. Scheible,P.A.,E.A.Rhoades,and M. Rhoades. 1977. Localization ofsingle-chain interruptions in bacterio-phage T5 DNA. I. Electron microscopic studies. J. Virol. 23:725-736.

25. Scheible,P.A., andM.Rhoades. 1975.Heteroduplex

mappingof heat-resistant deletion mutants of bacterio-phage T5. J. Virol.15:1276-1280.

26.Tessman,I. 1968.Mutagenic treatmentof double- and

single-strandedDNAphagesT4and S13 with hydrox-ylamine. Virology35:330-333.

27. Thomas,C.A., Jr.,andI.Rubenstein.1964.The ar-rangements of nucleotide sequence in T2 and T5

bac-teriophageDNAmolecules.Biophys. J. 4:93-106.

VOL. 29,1979

http://jvi.asm.org/