Inactivation of prophage λ repressor in Vivo

.I. Nol.

Biol.

(1979)

131, 553-572

Inactivation AI)RIANA

Lnhatoire

HAJLOKE,

of Prophage A Repressor in Vivo ALAIN

LEVINE

AXD RAYMOND

I)EVORET?.

Xection de Radiobiologie Ceblulaire d ‘Emynologie du C.X. R.S., 91.190 (rif-swr-

Ywtte

Prance

(Recrivrd

1 August

1.978, and

in rcwised form

26 Fdruary

1979)

.Jacob Ji Monad (1961) postulated that prophago A indllctiorl results from the, inactivation of the h repressor by a cellular inducer. Although it has been show~l t,hat the phage X repressor is inactivated by the recA gene product in vitro (Roberts inducer” in. &uu. et al., 1978), we wanted to determine the action of the “cellular Our results have led to a new model, which defines the relationship between t’llca “cellular inducer” and the recA gene product. In order t)o q~antitate the action of the celhllar inducer on the h repressor, wv(. tnnd~ use of bact,cria with elevated cellular levels of tlie X repressor (hyperimmune lysogens). We determined the kinetics of repressor inactivation promoted hy three represeutativc inducing treatments: ultraviolet ligllt irradiation, Ulymine deprivation and ternperature shift-up of t$‘:;f-l mutants. The kinetics of repressor decay in wild-type monolysogens iudicate tlrai repressor illactivation is a relatively slow cellular process that takes a generatio,) t.imr t,o rrach completion. Incomplete inactivation of the repressor without subsequent prophage development may occur in a cell. 1%‘:~call this phenomenon tlctccted at thcl biochemical level “subind~lctiorr”. lri ligperimnn~ne lysogens. subinductiori is always tire case. ,A Iligh crllular level of X repressor that prevents prophag? h induction does Ilot pre\-rnt, il~duction of a heteroimmune prophage such as 434 or 80. Although the caellular indncr~r does not seem specific for any inducible prophage, it does Ilot ~nact,i\-att, t,wo prophage repressors present in a cell in a random manner. We havch c*allctl tllis finding “preferential repressor inactivation”. Preferential repressor inactivatioli may be accounted for by considerin, 0 that bhe intracellular concon tr;lt iotl of a rqressor determines its susceptibility to the action of tile inducer. In bacteria with varying repressor levels, a fixed amount of repressor molecules is inactivated per urlit of tirne irrespective of the initial repressor concentratioll. The r&e of repressor inactivation depends on the catalytic capacity of the c*cbllular irlduccar that behaves as a saturated enzyme. In wild-type bacteria ttlr caellular irrtlncrr serlns to be produced in a limited amount, to 1lal.e a weak catalytic c:apacit,y a,1~1 a relatively short half-life. The amount’ of the inducer formed aft.cr tif- 7 rxpressiorl is increased in STS bacteria overproduciug a tij-l-modified ltcc.4 protein. Ttlis result is an indication that a modified form of t)lle RecA protein ca11ses r’(~pr~‘ss~~r illactivation %n 7%~. Front t’tlc%res~llts obtained WP propose a rnod~~l cboncerning the forrnatiori of thts c*rllular irlducer. 1Vt: postulate that the cellular inducer is formed iI1 a two-step rract,iorl. T11c is model visualises how the RecA pro&in can be induced to higlr cellular c:ollcrtltrations, even though the RecAp protoase molecules remain at a low c,orlcrrlt,ratiotr. The latter accounts for the lirnited protrolytic activity found in ?iW).

W22-2836/7Y/lW553-20

$02.00/O

~(0 1979 .&ademic

Press Inc. (London)

Lid

554

A. BAILONE,

A. LEVINE

AND

DEVORET

R.

1. Introduction Development of prophage h into vegetative phage (prophagt> induction) results from the inactivation of the repressor (Shinagawa & ltoh, 1973) : the disappearance of the active form of the h repressor is att,ributed to pdcolgtic cleavage (Roberts & Robert,s, 1975). Proteolytic cleavage of the A repressor can be produced in vitro with crude cell extracts of STS bacteria (Robe& et ul.: 1977) and wit’h relatively purified Red protein (Roberts et aZ., 1978). It has been shown in the past that prophage induction triggered by various agents or treatments is an indirect process (Borek & Ryan. 1958; Devoret & George, 1967: George & Devoret, 1971; Rosner et al., 1968). Inducing agents or treat#ments do not act directly on the h repressor but lead to the format,ion of a cellular inducer through a, complex cellular pathway whose end product is the cleaved repressor. Apart from prophage induction, inducing agents or treatments produce many other cellular manifestations (SOS phenomena), such as cell filammtation, induced phage repair, and phage and bacterial mutagenesis. Witkin (for a review, see Witkin, 1976) has postulated that all these cellular manifestations result from t’he inactivation of the various repressors that normally prevent them from occurring in an untreated cell. Among the SOS cellular manifestabions, prophage induction is known best at the molecular level and consequently its mechanism is one of the most’ amenable t,o molecular investigation. ln order to quantitate the action of the cellular inducer on the h repressor irk ~ivo. we made use of hyperimmune lysogens (Bailone & Devoret, 1978); t,hat is, bacteria with elevated cellular levels of the h repressor. We detJerminecl thr kinetics of repressor inactivation promoted by three representative treatments : u.v.t-irradiation, thyminc, deprivation and temperature shift-up of tif-1 mutants. From the results obtained we propose a model concerning the nature and the formation of the cellular inducer.

2. Materials and Methods (a) Bacterial The bacterial

strains

used in this work (b) Construction,

strains

are listed

in Table

of hyperimmune

1.

xtrak

sinctb Bacteria carrying plasmids encoding tllc hcI+ gene arc called hyperimmunr: hvir and hcI9OcY17 phages are not able to multiply in sucll host cells. Hyperimmunity is accounted for by elevated cellular levels of repressor (Bailone & Devoret, 1978). Elevated cellular levels of 434 repressor were obtained in bacteria carrying plasmid pGY 10 I (Levine et al., 1979) ; as expected, these bacteria are hyperimmune. They do not support, the growth of a virulent mutant of phage 434 giren to us by V. Pirrotta. For brevity, wf‘ to t,hat, designate bacteria carrying a plasmid and a prophage. whose CT gene is identical of the plasmid, as hyperimmune lysogens. (i) Transfer

of plasmids to lysogenTCs In practice, about IO6 I!-Zac+ The transfer was performed using I?‘-Zac+ sexduction. plasmid carriers and lo7 F- recipients were crossed in 10 ml of LB medium and incubated overnight. The transferred plasmid was selected by its antibiotic resistance marker: tetracycline (10 pg/ml) for pKB252 and kanamycirl (20 &ml) for pKB 103 or pGYlO1. The donor was counterselected by ampicillin or streptomycin. t Abbreviation

used : u.v., ultraviolet

light,.

1NaC;TlVATION

OF PROPHAGE

strains

I)11111x7 I)M1411 1)Ml 187 (Xi,!/WWi) L%oo (hinm-203) 294 (pKB252) GYlRX GYIOOX-I GY31.54 (:Y3155 GY31C1 GY3165 GY3180 (:Y3181 (:Y:{184

spr- 2 I lexd 3 tif - I 8fiA 11 sp-51 lerA3 tif-1 sfi-a 11 rec:l .upr-51 lerA3 t+f-1 sfiAl1 (hinm-L’Of) ZPX‘4+ rec.4 + (hinm-203) (pKB252) thy.46 th?yAG (A) UWA 16 hf. I (A) (pKB202) 1rvrAl6 &f-l (h) wrAlG thyA (X) (pKB202) uarAl6 thyA 63 (A) WTA 16 (Aimm434) (pKlW2) trorA 16 (kimm434) /crrAlG (hcI857)

(:Y:Llx!)

sp-bl lexA3 tif-I SfiAl I rocrl (pKB252) spr-51 lerA3 t;f-1 sfiAl1 (hind-) spr-51 lexA3 *if-l sfL411 (Xinrn-203) (pKB103) ~rzw.426 (X)2 /rwAlG (h)’ uarAlG (h) (Xh80) ~~~416 (A) ($80) rrmpd (A) urnpA (X%mm434) tmrA16 thyAG2 (aI)’ uvrA16 t;f-1 (X)l /~~r~416 (pRB202) ~~~arA16thyAG3 (pKB202) UWA 16 &f-l (pKB202) (Ximm434T) (pKB26”) (himm434T) spr-Al le.cA3 tif- I sfid 11 (X&m-203) rmpd (A) (pGYlO1) ampA (Ximm434) (pGY101) spr-51 lead3 tiff-l sfiAl1 (h&m-203) (pKB252)

(‘Y’IHW T . . (‘Y’~(i’Vl r . .. (:Y:w42 (:Y3643 (:Y:%46 GY3647 GY365’ (:Y3659 (:Y4xlfi GY4823 (:Y4834 (kY4842 GY4844 GYAGOB (:Ytitlll (iYtj614 (:Y(iHlti

- : ,thh,

1

TABLE

Bacterial

h KEl’HE880K

blount (1977) (:ift, from I). hlounl Roberts rl trl. (1977) (iift from D. Mount, Backman et rll. (1976) Httilonc ct frl. (1975) GY15X GY4834 GY3154 GY4823 GY3l63 GY481ti GY3180 (:y:j652 curc+d of A maths lyaogenic for Xc1857 I>M1411 11511187 1)M 1187 (hiww-203) (hi?rnl-2O:l) I)IIIllXi DM1187 (hinm-203) l)Ml187 (Airtm-203) I)M1187 (&m-203) Lcvinc et rtl. (1979) Levine at nl. (1979) GY3639 <. GYM39 Bailonc K- l)c:vorc~l~ ( 197R) GY4816 GY4816 294 (pKiH25”) GY4842 GY1199 . <* G Y364R QY3647

G I’6605

f’lasmitls ~ntt prophages uro indicated in parentheses. The expownt, tlesignetev whether thta Iysogen is a mono- or a dilysogen. The absence of exponent means that, t)he number of prophagn copies has not been measured. The sourre of each strain is indicat4, unless it WRS conatructecl specifically for this study.

(ii) Iutrotlrtction,

of a

homoimmune

prophage

into

h,yperimmune

Prophage X w-as transferred into an F- StrR (pKB202) cxrry iny a ChlR mutation located close to altX site, and l)irrallt’s (Xdhya et al., 1968) \?;a,~ done in a BBL Gaspak propl\age h and plasmid pKB202 among the recornbinants spont,aneous phage production. The phage to bacterium plasmid carriers with less than 2 to 50/ segregants or 10s3 lost the plasmid.

bacteria

recipient from an HfrH, alsc~ selection for ChlR St@ reconsystem. The presence of both was assessed by the level of ratio was low5 in cultures of in cultures of bacteria having

556

A. BAILONE,

(iii) Quick test for

h hyperimmunity

A. LEVINE

AND

R. DEVORET

The degree of hyperimmunity is defined by the patter11 of growtll of ultra\irulent, mutants of phage X (Bailono & Dcvoret, 1978). In practice, streaks of a phage stock at, about lOlo phage/ml were crossed against streaks of bacteria at IO9 bacteria/ml on LA plates. After overnight incubation pKB103 and pKB202 carriers could grow in contact with either XcI9OcY17 or h12 but not with h146. Bacteria carrying pKB252 plasmid could grow in contact with X146 but not with A668. (iv)

Lysogenisation

of hyperimmune

bacteria with a heteroimmune phage

About lo6 himm434 or himm434T phages were spotted on a lawn of hyperimmune bacteria carrying either the pKB202 plasmid or the pKB252 plasmid. Bacteria from the turbid phage spot were grown in LB and streaked on LA plates. The resulting colonies were tested for X immunity conferred by the plasmid and for 434 immunity conferred by the prophage. Since pKB.202 carried a large fragment of the A genome including the immunity region (see Table 2), pKB202 Ximm434 lysogens were also detected by the release of XimmX recombinant phages among Aimm434 parental phage progeny.

(c) Segregation

of plasmids

Plasmid-less derivatives were isolated as segregants by checking colonies for loss of the antibiotic resistance marker carried on the plasmid or for loss of the immunity to colicirl El. In the latter case, plasmid-less colonies were identified by replica plating on LA plates overlayered with soft agar containing 0.1 ml of a colicin El stock. Plasmid-less derivatives had lost, all the properties conferred hy the plasmid in question (see Table 2). (d) 1’hage straiw The phage

strains

used in this work are described it1 the accompanying paper (Levirw for the ultravirulent phage mutants h146 and h668 (Bailone & Dcvoret. 1978) and phage Xinm-203 (Roberts et al., 1977). Phage h12 is our stock of the classical Avir (Jacob & Wollman, 1954).

et al., 1979), except

TABLE 2 Properties

conferred to host bacteria by various

plasmids

Pldsmids Properties

to X to 434

Tetracycline Immunity

pKB20d

pKB25Z

t

Hyperimmunity Hyperimmunity Kanamycin

pKB103

of host bacteria

A I

resistance

I

resistanct

I

to colicin El

Phage X production lysogens Phage 434 production lysogens

pGYIO1

I-

1.

/

in reduced

retlucetl

rctlucrtl

normal

normal

normal

!lOITlld

reduced

+

i-

--

1

in

Recombinational exchange c1 gene in heteroimmune lysogens

of

These properties have been described by Backman (1979), Bailone & Devoret (1978) and in this work.

(1977), Backman

et al. (1976), Levine

et ~2.

INACTIVATION

OF PROPHAGE

(e) Cultures

and

h REPRESSOR

media

(i) Cultures Unless stated otherwise, bacteria were grown at 37°C. For the preparation t,xtracts, bacteria were harvested in the mid-exponential phase of prowt#h.

of crudcl cell

Synt,h&ic media were: YM9 (11 g Na,HPO,*iH,O, 3 g KH,PO,, 1 p NH,Cl, I 1 bidistilled water) ; YMSA (10 g Casamino acid, vitamin-free Difco decolourized with charc.oal, 2 g glucose, 0.2 g thiamine, 1 1 YM9); M63 (Cohen & Rickenberg, 1955) ; and MBBA ( 10 g Casamino acid, vitamin-free Difco decolourized with charcoa,l, 2 g glucose, 0.2 g thiamine, I 1 M63). (2)rnplet.e media were: BT (5 g NaCl, 5 g Difco tryptone, 8 g peptonc. I I demirlrraliz~d water) and LB (10 g NaCl, 10 g Bacto-tryptonc, 5 g yeast extract, 1 1 demineralized M.at,(lr). ( :‘r arid LA media were, respectively, BT and LB containing 15 p Biomar agar.‘l. (f) Inducing ( i) Th,ymine

treatmen,ts

depriuatiorL

Exponentially growing bacteria in YMSA supplemented with 40 pg thymine/ml were lrarvested by centrifugation, washed twice in cold YM9 medium, resuspended in prewarmed YMHA medium at about 5 x lo7 cells/ml and incubated at 37°C. (ii) Ultraviolet

light irradiation

Cells growing exponentially in BT or in LB medium were cent,rifuged, resuspended in The thickness of the cell suspension wvas 0.01 >I-MgSO, and exposed to u.v.-irradiation. less t,han 1 mm: 1i.v. doses were measured with a Latarjet dosimetar. Pllotoreact.iv}ltioll ivas avoided. (iii) tif- I -,wrovboted

induction

A culture of tif-1 bacteria growing exponentially at 32’C it, cithrr YMSA or Mfi3A Inedillm was shifted to 42°C and supplemented with 100 pg adenine/ml. Expression of ir:f’-l in a spr-54 @All genetic background could also be promoted by a temperatllrc\ sllift in rich LB medium. (g) Quantitation

of repressor

decay

Sodium azidc (5 mM) stopped repressor inactivation in bacterial cultures, which wcrv then frozen quickly. Sonicated cell crude extracts were prepared and quantitation of repressor decay using the DNA binding assay for repressor activity (see Levine et al.. 1979) was done as follows. Varying amounts of cell crude extract were added t)o react,ioll lnixtures in which the concentration of “operator DNA” was fixed and largely in excess over that of repressor molecules. The cell extract concentration in the binding ass+ Increased in proport.ion to the repressor disappearance, so that t,he binding of the decaying repressor t,o operator DNA was measurable. To ensure that other DNA-binding proteins did not impair the assay for repressor activity, we performed a reconstruction experimtrnt in wllich the repressor from a X lysogen was diluted int)o increasing amounts of a ~11 cbxtract of a non-lysogen (Fig. 1). It can be seen that over a large range of cell extract concentrations the cellular DNA binding proteins were not in such a quantity as to i rrt,erfere in the repressor assay. .4rr example of quantitation of repressor decay in thyA lysogens deprived of thyminc: is given in Fig. 2. In this experiment a series of homothetic curves were obtained, all of \\-llicll can be superimposed by changing the scale of the abscissa. This family of curves obeys a linear relationship. The slope value of each curve is proportional t.o the repressor concentration in the corresponding crude extract. Repressor decay was expressed as t(llc ratio [R]/[R,], [R,] being th e repressor concentration in the crude extract of the nont,rcated control lysogen. [R]/[R,] is the fraction of repressor still active after an inducing treatlnrnt. The value of repressor inactivation was defined as I - [R]/[.R,].

558

A. BAILONE,

A. LEVINE

AND

R. DEVORET

oA ,I0 1 5

do

00 I 95

1

I 145

Protein odded

I

1 195

(pg )

Fro. 1. Effect of increasing amounts of cellular proteins on the detection of repressor-DNA complex. To standard reaction mixtures (Levine et al., 1979) containing a cell extract from GY4823, R pKB202 carrier, (45 pg protein/filter) and increasing amounts of a cell extract from the nonlysogenic strain GY158, were added 3.5 pg of h or Xim~434 DNA. The amounts of h DNA (a) and Aimm434 DNA (0) retained on filters and the difference between these values (x ) (“operator DNA”) are plotted on the ordinate; the values plotted on the abscissa are the sums of the amounts of the cell extracts deposited per filter.

a i5

Protein

added

(pg )

Pro. 2. Quantitation of repressor decay in thymine-deprived lysogens. (a) The reaction mixtures contained 4 pg of X or Ximm434 DNA and increasing amounts of cell extracts prepared from GYlOOS-1 lysogens deprived of thymine for 0 min (O,.); 26 min (n,A); 50 min (V,V); 100 min (0,m); and 160 min (O,+). The amounts of h DNA (closed symbols) and hiwwa434 DNA (open symbols) retained on filters are plotted on the ordinate as a function of the amount of cell extracts in protein per filter indicated on the abscissa. (b) The data from (a) are expressed as A operator DNA specifically retained.

ISACTIVATION

OF

PROPHAGE

X REPRESSOR

.i.iO

3. Results (a) Kinetics

of repressor decay upon iru%uctiorL of X lysogens

The activity of the h repressor disappears in lysogens exposed t,o such inducing C (Shinagawa bi Itoh. treatments as y-irradiation (West et al., 1975) or mitomycin 1953). Since titration of the X repressor had to be performed on a large cell population. the question arises of what is the kinetics of repressor decay in individual cells as compared to the kinetics of repressor decay in the average cell population? One of two alternative situations might exist, depending on the action of the cellular inducer. .Hypothesis (1) : The cellular inducer inactivates quickly and at a random time all the rc’pressor molecules in an individual cell: the proportion of repressor left ill t ho (41 population would correspond to the frequency of non-induced cells, which would c*ontein a normal amount of h repressor. The kinetics of repressor inact’iva.tion would t hrrefore be identical t#o the kinetics of production of infective centres. Hypot#hesis (2): The cellular inducer takes a long time to inactivate the repressors molecules in an individual cell ; one expects t,hen a lag between t,he beginning of rrpressor decay and t,he onset of phage development,. The lag would measure thrb period required for the inactivation of virtuallv all the repressor molecules present in il lysogenic cell before the inducing treatment. Thn kinebics of rcapressor inactivation and of prophage development were measnr(~tt ~irllult,a.neon~l~ in order to differentiate hypothesis (I) from (2). (i)

Kepwssor

disuppmrwrce

in lysogens

deprived

of thym iw

Th(b proportion of repressor remaining in a bact’erial populabion was determinr4 bg the repressor binding assay; the proportion of derepressed cells was measured by the produet~ion of infective centres and free phage (Pip. 3).

7 I

I

.

-C

90 Time (mm) lk.

3. Repressor

decay

and prophage

induction

in thymine-starved

lysogens.

Bacteria GY3652 were incubated in thymineless medium. At t,he times indicated on the abscissa, the h repressor was titrated in cell crude extracts. The decay of repressor (A) is indicated on t,he left ordinate. In a parallel culture of GY3662 the infective centres produced (A) and the free phege released (0) were determined as a function of the thymine deprivation periods indicated on the abscissa. Thymine was restored for 2 h in the cultures in which the release of free phages was measured. Infective centres are plotted on the right ordinate as the fraction of the maximal viable cell count (1.6 x 10s at 10 min of thymine starvation) and free phages as the fraction of the maximal phage yield (8 x lo7 at 80 min of thymine deprivation).

560

A.

BAILONE,

A.

LEVINE

AND

R.

DEVORET

After thymine removal, repressor inactivation was almost immediate, whereas a lag of about 40 minutes was found between the beginning of repressor decay and the formation of infective centres. Prophage derepression occurred only when 90% of the repressor had been inactivated. The rate of repressor decay in the bacterial population seems to reflect the rate of repressor deca,y in individual lysogenic cells deprived of t,hymine. As a control we determined that in t,wo non-inducible recA128 and inf-3 mutant lysogens there was no decay of repressor activity after two hours of t,hymine deprivation (data not shown). This readily accounts for their failure to undergo lysogenic induction after thymine deprivation (Devoret & Blanco, 1970; Bailone et al., 1975). (ii) Repressor disappearance

in u.v.-irradiated

lysogens

Bacteria were exposed to increasing U.V. doses. Repressor inactivation was measured as a function of the time of incubation in rich medium after irradiation. The lysogens were uvrA so as to prevent excision repair of the pyrimidine dimers formed in the DKA. Repressor inactivation (Fig. 4) reached a plateau value at 30 minutes whatever the

0 Time (min) FIG. 4. Kinetics of inactivation of X repressor in cells exposed at increasing U.V. doses. Bacteria GY3639 exposed to U.V. at doses of 0.5 (a), 1 (v ), 3 (0) or 25 (‘I ) J/m2 were incubated in BT medium. At the times indicated on the abscissa, the h repressor was titrated in calculated as described in Materials and cell crude extracts. Inactivation of the /\ repressor, Methods, is plotted on the ordinate.

U.V. dose applied; this indicated that the number of pyrimidine dimers determines the rate of repressor inactivation in the population of u.v.-irradiated cells. The rate of repressor inactivation found is only an apparent rate. Repressor inactivation may occur in either of two ways : (1) at a rate increasing with the amount of DNA damage in every u.v.-irradiated bacterium; (2) at a rate independent of the amount of DNA damage but in a fraction of cells increasing with the U.V. dose applied to the cell population. According to the second hypothesis it is expected that, within a range of U.V. doses that do not affect the cell capacity to reproduce X, the kinetics of inactivation of repressor will level off at plateau values corresponding to the frequency of induced lysogens. Indeed, a rough proportionality between the fraction of repressor inactivated and the fraction of cells forming infective centres was observed (Fig. 5), except in cells exposed to a low U.V. dose (0.5 J/m2). In this case, 30% of the repressor

INACTIVATIOS

OF PROPHAGE

X REPRESSOR

56 I

Dose (J/m’) Fm. 5. Repressor decay and prophage development as a function of the u.v. dose. Bacteria GY3639 were exposed to U.V. doses as indicated on the abscissa. Irradiat,ed cells were incubated for 30 min in BT medium before being assayed for h repressor (v ) whose decay ix indicated on the left ordinate. The infective centres formcsl (n), plotted on the right ordinate as the fract,ion of the original number of viable lysogens, were assayed using t,hr ampicillin met,hod described by Moreau et al. (1976).

was inactivated, whereas only a small proportion of cells (3%) produced an infectivtl centrct. The latter result indicates that incomplete inactivation of the repressor rna,y occur in a cell without subsequent prophage development. We call this phenomenon “subinduction”. Subinduction has been observed also in lysogens transient’ly deprived of th ymine . The temporal relationship between repressor inactivation and prophage derepression was determined in cells exposed to a U.V. dose (3 J/m2) that causes maximal prophage induction. The average time of appearance of progeny phage measured in a one-step growth experiment was 60 minutes at 37°C (Fig. 6). This period accounts for the time elapsed for prophage derepression and subsequent development of the derepressed prophage. The average time for the development of thermally induced prophage Xc1857 was 30 minutes at 37°C (Fig. 6), the A~1857 repressor being inactivated in a few seconds at 42°C (Kourilsky et al., 1971). Therefore, the time for completion of prophage Cl+ derepression caused by u.v.-irradiat’ion is about 30 minutes. As observed in thymine-deprived cells, repressor inactivation after u.v.-irradiation is a relatively slow cellular process, it takes about a generation time to reach completion. (b) High cellular concentrations

of repressor preven,t prophage induction

In order to determine the effect of an elevated cellular level of the h repressor on prophage h induction, we used lysogens carrying the pKB202 plasmid, which increases fivefold the maintenance level of a monolysogen (Levine et d., 1979). When various hyperimmune lysogens were submitted to such inducing treatments as thymine deprivation (Fig. 7(b)) (Melechen et aZ., 1978), u.v.-irradiation (Fig. 10) or tij-l-promoted induction (Fig. 7(a)) little, if any, prophage development was observed. The proportion of infective centres produced, between 2y0 and 67;, corresponded to t’he frequency of cells t,hat’ had spont,aneously lost the pKB202 plnxmid

562

A. BAILONE,

A. LEVINE

AND

R. DEVORET

IO”

IO”

O- ., *R.,p-o-0/-.-c- 0 . 5 f .G IIx 2 .z / --I al t -I II . .-E za b bI f 30L -2- ,I I --2 E I , d d -o~o-.

-J-Y , , lm3 0

30

60

90

120

Time (min)

FIG. 6. Latent periods for prophage development in u.v.-irradiated hcI+ and h&357 lysogens. Bacteria GY3639 and GY3184 grown at 32°C in BT medium to lOa cells/ml were harvested by centrifugation, resuspended in 0.01 na-MgSO, and exposed to a u.v. dose of 3 J/m2. Bacteria were diluted lo-fold in BT medium prewarmed at 42”C, incubated at this temperature for 3 min, then diluted again IO-fold in BT medium and incubated at 37°C. At the times indicated on the abscissa, the cultures of GY3639 (0) and of GY3184 (0) were assayed for the production of plaque-forming particles (infective centres or free phages), plotted on the ordinate as a fraction of the maximal yield. The data on repressor inactivation in GY3639 (A) plotted on the ordinate are from t$he experiment described for Fig. 4.

(data not shown). Thymine deprivation followed by u.v.-irradiation produced a higher level of infective centres than either treatment when applied to a monolysogen, whereas such sequential treatment did not induce the prophage in a hyperimmune lysogen (Fig. 8). The lack of prophage induction in hyperimmune lysogens may result from an incomplete inactivation of the repressor due to a limited action of the cellular inducer. In order to quantitate the action of the cellular inducer, we measured the kinetics of repressor disappearance in u.v.-irradiated bacteria possessing increasing maintenance levels of X repressor (Fig. 9(a)). The concentration of repressor decreased roughly linearly with the time of incubation of irradiated cells. As compared to that of a monolysogen, the relative rate of repressor disappearance decreased by a factor proportional to the relative repressor concentration in non-treated lysogens : 1, 1.5, 5 (Fig. 9(a)). After u.v.-irradiation the amount of the inducer produced seems to be “saturated” by the repressor molecules, even in a monolysogen. The absolute rate of repressor inactivation is determined by the ca,talytic capacity of the inducer and is not influenced by the cellular repressor concentration. Note that in pKB.202 carriers repressor inactivation levelled off when the repress01 reached a value corresponding to about 2*5-fold the maintenance level in a monolysogen. This limited repressor inactivation accounts for the fact that there was no prophage induction in u.v.-irradiated hyperimmune lysogens (Fig. 10). Does limited repressor inactivation result from the occurrence of de novo repressor synthesis that would counterbalance repressor inactivation ‘1 To test this possibility we took advantage of the fact that tij-l-promoted repressor inactivation takes place in chloramphenicol-treated cells (West et al., 1975; Shinagawa et al., 1977), although protein synthesis is stopped. We found that treatment of tif-1 (pKB202) carriers with

LNACTIVATION

OF PROPHRGE

h REPRESSOR

503

IO” 2

I hY-

tif-l

-pKB202 I -

h A’ -pKB202 a**

v a’

f

/o-o

-21s J -3? I tpKB202

p

,~de,~pKB202

-46

,/O

0 /

’ 0

30

60

0 Time

, 30

60

90

(mini

(a)

(bl

Fro. 7. Absence of prophage induction in hyperimmune X lysogens submitted to various inducing t-recttmems. (a) Expression of @-I. Bacteria GY3154 and GY3166 grown at 32°C were incubated for 30 min in M63A medium supplemented with and without 100 pg chloramphenicol/ml and then shifted to 42°C by a lo-fold dilution of the cultures into the seme supplemented with adenine (100 W/ml). At the times indicated on the abscissa, chloramphenicol-treated (closed symbols) and untreated cultures (open symbols) were assayed for t,he production of infective cent~res scored after OVWnight incubation at 32% Infective centres produced by GY3154 (circles) and by GY3155 (triangles) bacteria are plotted on the ordinate as the fraction of the initial viable cell count, (10” cellshnl in t,he controls and 5 x IO6 cells/ml in the chloramphenicol-treated cultures). (b) Thymine deprivation. Bacteria were deprived of thymine for the periods of time indicated 011 the abscissa. The number of infective centres formed by GY3163 (0) and by GY3165 (0) bacteria, is plotted on the ordinate as the fraction of the maximal viable cell count (respectively, 1.9 x 10’ cells/ml and 1.5 x lo6 cells/ml at 10 min of thyminr deprivation).

chloramphenicol did not affect the rate of repressor disappearance after expression of the tif-1 mutation (Fig. 9(b)). If the process of de novo repressor synthesis efficiently counteracted the action of the cellular inducer by preventing de n,ovo repressor synthesis, we should have found an acceleration of the kinetics of repressor-disappearance but this was not the case. In a monolysogen the cellular repressor was inactivated within 20 minutes after expression of the tif-I mutation. In contrast, in pKB202 carriers, after a lag of 15 minutes, the repressor concentration decayed linearly with time within the first 60 minutes down t,o a tailing-off value, half of that of the maintenance repressor level in a monolysogen (Fig. 9(b)), high enough to maintain the prophage in the repressed st,at’e (Fig. 7(a)). (c) Induction

of a heteroimmune concentrations

prophage

in cells with

high cellular

of the X repressor

Each lambdoid phage such as A, 434, 21 or 80 displays a specific immunity determined by a repressor protein that recognises a specific operator sequence. Yet, all inducible prophages respond equally to inducing agents or treatments. Therefore, the

564

A. BAILONE,

A. LEVINE

0

60

AND

0

120 Time

60

12. DEVORET

120

180

(min)

FIQ. 8. Absence of prophage induction in hyperimmune A lysogens deprived of thymine and exposed to U.V. Bacteria GY3163 and GY3165 were incubated in thymine-less medium. At the times indicated on the abscissa, one part of the culture was exposed to u.v. (2 J/m2 for GY3163 and 0.8 J/m2 for GY3165) the other part being kept unexposed. u.v.-irradiated (triangles) and unexposed cultures (circles) were assayed for viable cells (filled symbols) and for infective centres (open symbols).

cellular inducer, the ultimate product of the inducing treatments, must be nonspecific. In a cell containing the h and the $80 repressors, one would expect the two repressors to be inactivated with a probability corresponding to their relative cellular concentrations if the inducer does inactivate repressors at random. In such a heteroimmune dilysogen, the presence of the 480 repressor should affect the kinetics of A repressor inactivation, since a limited number of repressor molecules are inactivated (see section (b), above). The results obtained are in disagreement, with the above hypothesis, since the 480 repressor did not influence the kinetics of X repressor inactivation in u.v.-irradiated heteroimmune dilysogens (Fig. Q(a)). In order to investigate further whether a high level of X repressor would prevent induction of a heteroimmune prophage, we introduced into hyperimmune bacteria (carrying plasmid pKB202 or pKB252) prophage 480 or Ximm434. The himm434 (Fig. 10) or the 480 prophage (data not shown) were induced at the same U.V. dose and with the same kinetics as in the control lysogen having lost the plasmid. Similarly, induction of prophage h (Fig. 11) but not of prophage himm434 occurred in cells carrying plasmid pGYlO1 that amplifies the 434 repressor level. The repressor of a lambdoid prophage, even at an elevated cellular concentration, does not compete with the repressor of another lambdoid prophage in its interaction with the cellular inducer. (d) Increased repressor inactivation.

in, STS bacteria

Genetic and biochemical data have shown that the tif-1 mutation lies in the recA gene (Morand et al., 1977; Castellazzi et al., 1977) and modifies the recA gene product

INACTIVATION

0

1

OF PROPHAGE

I

I

I

30

60

90

1

1

120 150

1 0

0

8

30

60

566

X REPRESSOR

Time (min)

FIG.

9. Kinetics

of repressor

(cl

(b)

(a)

disappearance

in hypcrimmune

bacteria.

(a) After u.v.-irradiation. Cells were exposed t,o a u.v. dose of 10 J/m2 and incubated at 32% in RT medium. At times indicated on the abscissa the h repressor was titrated in crude cell extracts. The decay of repressor in GY3639 (A), GY3638 (a). GY3643 (V), GY3642 (x) and in GY4816 ( n ) is plotted on the ordinate. The initial concentrations of X repressor (units per pg prot,ein) were: (A) 0.8; (r) 0.8; (0) 1.2; (x) 1.2; ( n ) 3.9. (b) After tij-I expression. Cultures of tq-1 bacteria grown at 32°C in M63A medium were diluted s-fold in M63A medium prewarmed at 42°C and supplemented with 100 pg adeninelml. Cells were Incubated at this temperature for the periods indicat,ed on the abscissa and the h repressor wus titrated in crude cell extracts. The h repressor decay in GY3659 (A), GY4834 (0) and in GY4834 treated with chloramphenicol (0) is plotted on the ordinate. In the chloramphenicol-treated culture the drug (100 pg/ml) was added 15 min before the temperature shift-up. There was no rlecay of repressor activity in GY4834 treated with chloramphenicol and incubated for 2 h at, :iZ”(‘. The initial concent,rations of h repressor (units per pg protein) were: (A) 1.3; (0) 6.0: (’ I) 5..s. (c) After tif-1 expression in an STS genetic background. Cells grown in YMSA medium at 32’C’ were shifted to 42°C by a 2.fold dilution in YMSA medium prewarmed at 60°C and supplemented \\ ith 100 pg adenine/ml. Cells were incubated at 42°C for the periods indicated on the abscissa and the A repressor was titrated in cell crude extracts. The decay of repressor in GY6605 (a), GY3199 There was no decay of repressor activity in (‘I) and in GY6616 (A) is plotted on the ordinate. GY3189:STS recA (pKB262) within 2 h after the t,emperature shift-up. The initial concentrations of’ X rc~prvssor (units per pg protein) wese: (0) 0.9; (7) 2.7 (A) 20.

(McEntee,

1977 : Gudas & Mount, 1977) ; the Reed protein is overproduced in STS : spr-51 tif-1 sjAl1 bacteria (Gudas & Mount,, 1977). We wanted to determine whether the overproduction of the reed tif-.l gene product increases the action of the cellular inducer on the h repressor in bacteria with elevated rcapressor levels. With this aim we introduced either the pKBlO3 or the pKB252 plasmid into STS (Ainm-203) lysogens. Phage hinm-203 was used, since it can produce stable STS lysogens at 32°C. The km-203 mutation may generate a new promoter t’hat results in an increase of the rate of repressor synthesis (Roberts et al.. 1977); in effect, we found that the repressor level in STS (hi,nm-203) was strongly reduced by treatment with chloramphenicol (Table 3) or kanamycin (data not shown). There seems to be a dynamic equilibrium between inactivation and repressor synthesis in STS (h&m-203) lysogens grown at 32°C. The spr-51 mutation increased the rate of repressor inactivation and the amount of repressor inactivated after tif-1 expression (Fig. 9(c)). The plasmids pKB103 and

566

A. BAILONE,

A. LEVINE

AND

R. DEVORET

h 2

:

-6 t0

3

9c

6 Dose (J/m’)

FIO. 10. Ultraviolet plasmids.

light

induction

of the 434 prophage

in cells

carrying

pKB202

or pKB252

Bacteria grown in LB broth to 10s cells/ml were harvested by centrifugation and resuspended in YM9 medium. This suspension was diluted serially in the same medium to cover 6 concentrations ranging from non-diluted to IO- 5. A 2.ml sample from each dilution was placed in a 5.cm glass Petri dish. The 6 dishes were u.v.-irradiated all together and given cumulative doses. For each U.V. dose the numbers of infective centres (open symbols) and of viable cells (closed symbols) were measured in t)he following bacteria: GY3163 and GY3165 (left panel, circles X, triangles X, pKB202); GY3180 and GY3181 (middle panel, circles 434, triangles 434, pKB202); GY4842 and GY4844 (right panel, circles 434T, triangles 434T, pKB252).

20

40 Dose

FIG.

11. Ultraviolet

light-induction

0

20

40

( J/m’)

of h prophage

in cells carrying

plasmid

pGY101.

Growth of bacteria and u.v.-irradiation were done as described for Fig. 8. For each U.Y. dove the numbers of infective centres (open symbols) and of viable cells (closed symbols) were measured in the following bacteria: GY6614 and GY3647 (left panel, circles 434, triangles 434, pGY101); GY6611 and GY3646 (right panel, circles h, triangles X, pGY101). Triangles and circles designate lysogens carrying plasmids and t)heir plasmid-less derivatives,:respectively.

1 NACTIVATION

OF PROPHAGE TABLE

E#ect of chlorumphenicol

(:Y6605:

3

on the repressor level in. lvTS (hinm-203)

STS(Xi,/m-203)

367

h REPRESSOR

lysogens

32

0.8

42 32

-co.1 0.2

4”

_ :0.1

32

I.0

42 32

0.i 1.0

42

0.7

3%

1.2

42 :I?

1.3 1.:I

42

I.1

+

GY3186:STS

(hi,&)

-

+

C600 (Aim-203)

-t

(klls were grown in YMSA medium. The concentration of chlorumphenicol (CAM) was ZOOpg/ ml. Cells were incubated for either 30 min at 32°C or 20 min at 42°C. Incubation at 42°C was done in medium supplemented with 100 pg adenine/ml. The repressor units indicated in column 3 are as in Levine et nl. (1979).

pKB252 increased the level of repressor in the STS (Xi~zm-203) lysogen threefold and 20-fold, respectively. In such bacteria the rates of repressor inactivation augmented in proportion to the initial repressor concentration (Fig. 9(c)). This result indicates that t)here was more inducer formed in the STS mutant than that formed in wild type and t
568

A. BAILONE,

A. LEVINE

AND

R. DEVORET

DM 1187 extract (~4) FIG. 12. Inactivat,ion of the /\ repressor i)7 vitro. DM1187 bacteria were grown in BT medium at 32”C, wheroas 294 (pKB252) bacteria were grown in LB medium supplemented with 10 pg tetracycline/ml at 37°C. Cell extracts were prepared by lysozyme lysis (Roberts et al., 1977). The reaction mixture for repressor inactivation in vitro was as described by Roberts et al. (1977), except that MgClz was 5 mM. Each reaction mixture of 100 pl contained 177 pg of protein of a cell extract of 294 (pKB252) corresponding to 2220 units of /\ repressor and the amounts indicated on the abscissa of protein of a cell extract from DM1187. Samples of 10 ~1 taken before (0) and after (0) 20 min incubation at 39°C were included in a 600-pl reaction mixture for repressor assay containing 11 pg of either X or hinam434 DNAs. The DNAs retained values on the ordinate are the differences between t,he amounts of A and &mm434 for 200 ~1 of the DNA binding mixture deposited on filters.

4. Discussion (a) Inactivation

of the A repressor

and prophage

developmed

Upon inducing treatments the repressor decays with time in individual cells with relatively slow kinetics. Prophage development occurs when the repressor level in cells is below a threshold value estimated to be about lOo,b of the cellular concentration in non-treated lysogens. Incomplet,e inactiva,tion of the h repressor can occur in a cell and not lead to subsequent prophage development. This subinduction detected at the biochemical level is not seen at the microbiological level. In hyperimmune lysogens the process of repressor inactivation does not go to completion, so that there is no prophage induction. Subinduction is the rule. This fact suggests that in a wild-type-inducible lysogen the cellular concentration of the prophage repressor is high enough to maintain the prophage state but not too high to prevent lysogenic induction from occurring.

(b) Action

of the cellular

inducer

upon

the h repressor

A high cellular level of X repressor that prevents prophage h induction does not prevent induction of a heteroimmune prophage such as 434 or 80. Although the cellular inducer does not seem specific for any inducible prophage, it does not inactivate two prophage repressors present in a cell in a random manner. We have called this unexpected finding “preferential repressor inactivation”. Preferential repressor inactivation may be accounted for by considering that the intracellular concentration of a repressor determines its susceptibility to the action of the inducer. The X repressor can exist as a monomeric or a dimeric molecule ; there is a concentration-dependent equilibrium between these two forms (Chadwick et al.,

IXACTIVATION

OF PROPHRGE

.;,A!)

A REPRESSOR

1970; 1~. Sauer & M. Ptashne, personal communication). Preferential repressot in&iv&ion may result from preferential inactivation of repressor monomers. This phenomenon can also be interpreted by assuming that t’here is a cellular comparbmentjat,ion of repressors when they are at a high concentrat,ion ; the susceptibilit: t.0 cleavage of a repressor might then depend on it’s cellular location.

(c) Co-or&ate

expression

of SOS phenomena

and inactivation

of repressors

.Igents and t,reatments that induce prophage X in Escherichia coli also produce ba.cterial mut,agenesis, reactivation and mutagenesis of u.v.- damaged phage h, and delay cell division. Witkin (1967,1976) suggested that these induced phenomena (now designa,ted SOS phenomena) result from the inact,iva,tion of specific repressors rwponding bo t,hr same inducer. .I. set of cellular functions to be co-ordinately induced would require that their rc~prrssors be all ina&ivated in a given time. Some split phenotypes can easily be wccoun&+d for as a result of (1) a mutation increasing t,he level of a given repressor swh that ifs inactivation may not be complete, and (2) a modification in the activit? or in the amount of t,he cellular inducer.

((1) The cellular

inducer

is an, activated,fraction

of thP RNA

prokin

.Iacoh & Monad (1961) postulated the existence of it cellular inducer t,hat would inactivate the h reprwsor. Our experimental results can be interpret’ed best by t#he h?rpothrsis that, t.he cellular inducer is produced in a limited amount: has a wr~tlc catalytic capacity, and has a relatively short half-life. The occurrence of proteolytic cleavage of the X repressor in eitrn as i/L vivo (Roberts $ Roberts, 1975; Roberts et al., 1977) suggests t’hat cleavage may be the mechanism by which the cellular inducer causes repressor inactivation. Among the genes rec14 (Brooks &I Clark, 1967; Hertman & Luria, 1967), ZexJ (Donch et nl.. 1970). i?/fil (Bailone et al., 1975) and recF (Horii & Clark. 1973; Robe&s $ Roberts. 1975: Armengod & Blanco, 1979) that cont’rol prophage X induction, which is the one coding for the protease? Genetic and biochemical data have shown that the tit-1 mutation lies in the recA gene (Morand et al., 1977: Castellazzi et al.. 1977: Me&tee, 1977). The s$ tif spr (STS) mutant, synthetizes the tif-l-modified form of t,he Reed protein at a high rate (Gudas & Mount, 1977) and displays an increased level of the cellular inducer produced after tif-I expression (this work). Extracts of STS bacteria contain an ATP-dependent protease that cleaves the h repressor in vitro (Roberts et al., 1977). This activity is due to the tif-modified RecA protein (Roberts et al., 1978). All these data suggest’ that the recA gene product cleaves the h repressor in vivo. The fee-4 gene product has been identified as that formerly called protein S (McEntee. 1977: Lit,tle & Kleid, 1977; Emmerson & West,. 1977; Gudas & Mount. 1977) synthetised in large amounts in cells submitted to an inducing treatment (Inouye & Pardee. 1970; Gudas & Pardee, 1976). Biochemical and genetical data point to bhe fact t,hat, the ZexA product controls recA expression as does racil itself (McEntw. 1979: Mount, 1977; Sedgwick et al., 1978). Bacteria carrying mutations lex.4 I or lex:d3 are unable to undergo prophage induction in spite-of the presence of a funct’ional recA gene product (genetic recombination being almost unaffected). These

570

A. BAILONE,

A. LEVINE

AND

R. DEVORET

results have been interpreted by the hypothesis tha,t recA derepression is required for lysogenic induction to occur (McEntee. 1977: Gudas $ Mount, 1977; Roberts et al., 1978). Amplification of the reed gene product rna,y not be necessary for the cellular inducer to be made active: in chloramphenicol-treated tif-2 bacteria the X repressor is inactivated and cleaved (Shinagnwa et al., 1977 : t,hix work), even though there is no amplification of the RecA protein (Maenhaut-Michel et al.. 1979). Conversely, amplification of the recA product by insertion of tJhe recA gene into a ColEl plasmid does not result in the induction of SOS phenomena, (McEntee. 1977). We propose that the RecA protein must be activated to act as a protease of prophage repressors. This process occurs in two steps, one of which causes the derepression of the recA gene (see Fig. 13). It is assumed that the rec;I and ZexA gene products

-

Induced

DNA repair

X repressor

Complex RecA - LexA

Effector

(ATPI

RecAa

I -

cleavage

RecAp

_/ infA -J

lesions

form a complex (Morand et al., 1977), which acts as a repressor of recA gene transcription. This assumption explains why ZexA- as well as recA- mutations prevent induction OfrecA gene transcription. As a consequence of the arrest of DNA replication, the complexed RecA protein combines with an effector and undergoes an allosteric change to become RecAa. A stalled replication fork may produce the effector or be the effector itself. The transformation of RecA into RecAa protein occurs in a tif-I mutant at 42°C in the absence of an effector. The RecAa protein has a reduced affinity for the LexA product and t.he RecA-LexA complex dissociates. The recA gene is derepressed. The RecAa protein must be activated by an activator, very likely ATP (Roberts et al., 1978), to act as a protease, which we call RecAp. The first reaction, RecA-+ RecAa, is negatively controlled by the EexA gene. The &A gene product acts with the RecA protein as a corepressor of the recA gene. Being an inhibitor of the formation of RecAa, the muk,tetl IPXA .- gent: product prevents thereby the activation of the RecA protein. The second reaction, RecAa-+ RecAp, is controlled negatively by the activated protease itself and positively by gene products such as infA. The X repressor may be

JNACTIVATION

OF PROPHAGE

h REPRESSOR

5il

Couturiw HE von Meyerburg. inactivattad at a st,nlled replication fork (Dewret,, unpublished rwultJs). The ;$4 gem: product may control the availability of the* activator at the site of the reaction or may stabilize the unstable RccAp protein. Our model for RecA activation into RecAp accounts also for the multiple role of the RccA protein in a, cell. When DNA damage occurs, a large amount of RecA must lw formed so as t.o serve as healing protein of the multiple DNA lesions. whereas its derivativr RecAp, a protease, is required only in a limit,& amount. We arc grat,eful t,o K. Backman alld M. Ptashne for strains and advice. We greatly tx~nefitrd from discussions with D. Mount and .J. Roberts and wt~ arc’ also thankful to I). Mount for the gift of strains. Thcx technical a,ssistance of M. Pierre is highly appreciated as wc,ll as that of D. Poircbt atld of J. Nappc. Euratorrr is acknowledged for providing grant 110. 223-76 BIOb’.

REFERENC:ES .4dhya., S., Cloary. P. & Campbell, A. (1968). Proc. Nat. dcad. Bci., r;.~S..4. 61. 956 Hti2. Armengod, M. E. & Blanco, M. (1978). Mutat. Rs~. 52, :S7-~47. Backman, K. (1977). Ph.D. thesis, Harvard University. l%ackman, K., Ptashne, M. & Gilbert, W. (1976). Proc. Xat. Acarl. Sci., I’.S.A. 73. 4174 4178. Bailorrc:, A. & De\-oret, R. (1978). c’irology, 84, 547%X0. Railone, A., Blanco, M. & Devoret, R. (1975). idol. Gen. Genet. 136, 291~-307. Borek, E. & Ryan, A. (1957). I-‘roc. Nat. Acad. Sei., C.S.A. 44, 374&377. Brooks, K. & Clark, A. ,J. (1967). J. I’irol. 1, 283-293. (!astellazzi. M., Morand, P., George, J. & Buttin. G. (1977). 1cloZ. Gevh Uenet. 183, 297 -310. (‘hadwick, P., Pirrotta, V., Steinberg, R., Hopkins, N. & Ptashne, M. (1970). Cold Spriq Harbor Sywq. f&ant. Biol. 35, 283-294. C’ohen. C. N. & Rickenberg, H. V. (1955). Ann. Inst. Pasteur, 91, 693S720. Devorut, It. & Blanco, M. (1970). Mol. Gen. Genet. 107, 272-280. Drroret, R. & George, ?J. (1967). Mutat. Res. 4, 713-734. Donell. J.. Greenberg, J. & Green, M. H. L. (1970). Gen,et. Res. 15. 87 -97. Ernmrrson. I’. T. & West, S. C. (1977). Mol. Gen. G&et. 155, 77-86. (icwrgt*, .I. & Drrorc%, R. (1971). Mol. Gerc. Genet. 111, 103 1 I!). (:ntlus. I,. .J. & Moutit. D. W. (1977). I’roc. Sat. had. Sci., ~T.A~.A. 74, 5280 52H4. (idas. I,. -1. 82 Par&r. A. B. (1976). J. illol. Hiol. 101, 459 477. Hc~rtmarr. I. & Louis, S. E. (1967). J. 1%1oZ.Bid. 23, 117 133. Horii, %. 1. & (Xark, A. J. (1973). ,/. .Ilol. Biol. 80, 327. 344. lnmyt~, M. & Pardee. A. B. (1970). .I. Biol. Ch>em. 245, 5813 *iSI!). .JILCO~, F. $ Monad. .J. (1961). -7. Mol. Bid. 3, 318 -356. ,Jarob, I<‘. & Wollmnn. E. I,. (1954). Anvc. Inst. Padew, 87, 653 --673. Kollrilsky, F’., Bollrguignon, M.-F. & Gras, F. (197 I). III The Bacterioph,age Lambda (Herstlcxy. .4. D., (1~1.). pp. 647WX6, C’oltl Hprillg Harbor l,ahtrratory, C:old Spring Hitrhor. I,c*vinc~. A.. Hailonc~. A. & Devoret’, R. (1979). ./. .\Zol. Hiol. 131. 055~WI. I,it#t.le. ,J. \\.. & Kleitl, I>. (4. (1977). J. Bid. Chew. 252, 6251-6252. Mitr.rltlwllt-Micll~~l. (i.. Hrwndrnburgcr. .L\. & BoiWux. S. ( I9iX). .I/o/. (/P)). (:p~t. 163, ‘“)‘j< -. .,(j(,* i* McEntcc. K. (1!175). f’roc. Sat. AC&. IS&., C.S.d. 74, 5275~5279. McEntw, I<. (1979). In Tj)n:A Repair Mechanisms (Hana~wult, P. (.‘. &, Frirdbrrg, E. ( ‘, chds). A(*adomic Press, Near- York. In tile press. Molechrn. N. I<:., Go, G. & Lozeron, H. A. (1978). !IfoZ. f:err. Co&et. 163, 213.-221. hlorarttl. I’.. (:oz(~. A. & Dcvoret. R. (1977). Jlol. (:en. (:~not. 157, 69 -82.

572

A. BAILONE,

A. LEVINE

AND

R. DEVORET

Moreau, I’., Bailone, A. & Devorct,, K. (1976). I’roc. Nat. Acad. Sci., I’.S.A. 73, 3700 3704. Mount, D. W. (1977). I’roc. Nat. ilcad. Sci., fi.S.A. 79, 300-304. Roberts, J. W. & Roberts, C. W. (1975). I’TOC. ,Vat. rlcacl. Sci., I7.N.d. 72, 147--151. Roberts, J. W., Roberts, C. W. & Mount, D. W. (1977). t’roc. Sat. Acad. Sci., U.S.A. 74, 2283-2287. Roberts, J. W., Roberts, C. W. $ Craig, N. L. (1978). I’roc. Xat. dcad. Sci., I T.g.il. 75, 4714-4718. Rosner, J. L., Kass, L. R. & Yarmolinskg, M. R. (1968). Cold Sprin,q Harbor Symp. @cant. Biol. 33, 785-789. Sedgwick, S. U., Levine, A. di Bailone, A. (1978). Xol. Gen. Genet. 160, 267- 276. Shinagawa, K. & Itoh, T. (1973). Mol. Gem. Genet. 126, 103-110. Shinagawa, H., Mizuuchi, K. I% Emmerson, P. T. (1977). Mol. Gen. Genet. 155, 87-92. West, S. C., Powell, K. A. t Emmerson, P. T. (1975). iVoZ. Gem Genet. 141, 1-8. Witkin, E. M. (1967). Proc. Nat. Ad. Sci., U.S.A. 57, 1275-1279. Witkin, E. M. (1976). Bacterial. Rev. 40, 869-907.