• Nenhum resultado encontrado

Résumé de l'article:

Des expériences de "footprint" de la région opérateur avec du répresseur de Mu purifié

ont permis de déterminer la taille de la région opérateur qui s'étend sur 184 pb (Krause et

Higgins, 1986). Elle contient trois séquences opérateur, 01, 02 (qui contient le promoteur

pE), et 03 (qui contient le promoteur pCM), qui contiennent respectivement trois, quatre,

et deux séquences consensus asymétriques de fixation du répresseur CTTTTN3W3 (Krause et

Higgins, 1986; Vogel et al., 1991).

L'approche in vivo classique d'isolement de mutants opérateur consiste à sélectionner

des mutants phagiques virulents, c/s-dominants, capables de surmonter l'immunité et de se

multiplier dans une bactérie lysogène. Dans le cas de Mu, elle a fourni de rares mutations

dans le répresseur (voir plus haut la description des mutations vir de Mu) mais pas dans

l'opérateur (Geuskens et al., 1991, 1992). Une explication possible de cette observation est

que les mutations de la région opérateur qui empêchent la fixation du répresseur, affectent

aussi systématiquement la fixation de la transposase et donc bloquent la réplication du phage

et sa capacité de former des plages de lyse.

Nous avons donc choisi une autre stratégie de sélection pour isoler des mutations dans

la région opérateur. Pour sélectionner des mutations qui permettent l'expression

constitutive à partir du promoteur pE, nous avons utilisé le plasmide pJV313, un plasmide à

haut nombre de copies qui exprime, à partir de son promoteur naturel pCM, le répresseur

galactosidase. Ce plasmide a été introduit dans une souche clpP de manière à augmenter la

répression de pE (Desmet, Geuskens et Gama, résultats non publiés). Cette souche n'exprime

la p-galactosidase qu'à partir de pE. En isolant des bactéries capables de se développer sur le

lactose comme seule source de carbone, vu le haut nombre de copies du plasmide pJV313,

nous nous attendions à n'isoler que des mutations c/s-dominantes, conférant le phénotype

Lac+. Les colonies isolées grâce à cette sélection répondaient effectivement à ce critère. Les

opérateurs mutés ont été séquencés ce qui a révélé la présence d'une parmi plusieurs

mutations ponctuelles ou d'une délétion.

La nature et la localisation de ces mutations, qui toutes affectent la région 02 ou ses

abords immédiats, suggèrent que

02

est le site le plus important soit dans la structuration

soit dans la stabilité du complexe opérateurs-répresseur. Les résultats suggèrent aussi

l'existence d'une hiérarchie dans les sites de fixation du répresseur localisés dans les

opérateurs 01 et 02. Enfin la caractérisation de ces mutations apporte de nouvelles lumières

sur l'importance relative des bases centrales (NNN) dans les contacts entre le répresseur et

la séquence consensus Cl i i i NNNWWW.

92

Article 4: Soumis à Genetics

In vivo Mutational analysis of bactériophage Mu

operators.

Lucie Desmet*, Marie-José Gama*, Jamal E. Laachouch*, loan Petrescu^

and Ariane Toussaint*»^

‘Laboratoire de Génétique des Procaryotes, Unité Transposition Bactérienne, Université

Libre de Bruxeiles, 65 rue des Chevaux, B1640 Rhode St Genèse, Belgium.

^ Laboratoire de Biochimie des Microorganismes, Université Joseph Fourier, BP 53,

F38041 Grenoble cedex, France.

Running head: Bactériophage Mu Operators

Keywords: phage Mu, repressor, operator

Corresponding author:

Ariane Toussaint

Laboratoire de Génétique des procaryotes

Université Libre de Bruxelles

65 rue des Chevaux

B1640 Rhode St Genèse, Belgium

Tel: 32 2 650 97 41

Fax: 32 2 650 97 44

ABSTRACT

In bacteria lysogénie for bactériophage Mu, the phage repressor binds to a tripartite

operator région, 01, 02, 03, to repress the lytic promoter pE, located in 02, and negatively

autoregulate its own synthesis at the pCM promoter located in 03. We isolated and

characterized operator mutations which lead to at least partial derepression of pE. Their

location reveals a hierarchy. between operator subsites and between the repressor binding

cts45 cts71 cts25 cts62

L I D Q

13 KSIWCSPQEIMAADGMPGSVAGVHYRANVQGWTKRKKEGVKGGKAVEYDVMSMPTKEREQVIAHLGLSTPDTGAQANEKQ 92

1 MELWVSPKECANLPGLPKTSAGVIYVAKKQGWQNRTRAGVKGGKAIEYNANSLPVEAKAALLLRQGEIETSLGYFEIARP 80

V7////////A yCHZ] V///////M

H3

W

[ YZZZ2t

B1 H1 H2

DNA binding

Figure 1. Alignement of the repressor and transposase N-terminal régions.

Residues 13 to 92 of Mu repressor hâve been aligned on transposase residues 1 to 80 using the GCG Bestfit program. The repressor cts

mutations (VOGEL et al., 1991) are shown at their positions. B, H, T, and W strand for b-sheet, a-helices, turn and wing respectively. The DNA

INTRODUCTION

Bactériophage Mu is a transposable temperate phage. Upon infection of a suitable host

(here E.coli K12), the injected viral DNA intégrâtes at random in the host genome. Although

this remains to be demonstrated, it is usually assumed that it is after this first intégration

event that the décision is made between remaining a silent, repressed prophage or entering

the lytic cycle. In the first case, repressor is expressed from the pCM promoter and binds a

tripartite operator, blocking transcription from the early promoter pE. Promoters pCM and

pE initiate overlapping transcripts in opposite directions, generating mRNA with

complementary 5' ends. In the second case, transcription from pE leads to the expression of

the transposition proteins pA (the transposase) and pB (a DNA dépendent ATPase which

favors inter-molecular transposition events) and other early lytic proteins. Viral DNA

copies accumulate at random positions in the host genome by successive rounds of réplicative

transposition (for review see SYMONDS étal. 1987, TOUSSAINT et al. 1994, HANIFORD and

CHACONAS 1992, MIZUUCHI 1992). The 250-bp Mu early regulatory région plays a

crucial rôle in both cases. In the prophage State this région is tightiy bound by repressor.

During réplication it is transiently bound by transposase to allow for the synapsis of the

first réplicative transposition intermediate (for review see HANIFORD and CHACONAS

1992, MIZUUCHI 1992) and hence is called lAS (for internai activating sequence). The

repressor and transposase N-terminal régions (around 70 amino acids) are homologous

(44% identity, HARSHEY et al. 1985) and define the operator binding moiety of the

proteins.

The operator région, as defined by footprinting experiments with purified repressor,

is 184 bp long (KRAUSE and HIGGINS 1986). It contains three operators, 01, 02 and 03

with respectively three, four and two asymmetrical

11

bp consensus repressor binding

sites (CTTTTN3W3, KRAUSE and HIGGINS 1986, VOGEL étal. 1991). It aiso contains the pE

(in 02) and pCM (in 03) promoters. Operators sites 01 and 02 overlap with the lAS

transposase binding sequence (LEUNG et al., 1989). Binding of Mu repressor to the

operators is stable and cooperative (VOGEL et al. 1991) and is facilitated by the host IHF

(GAMA étal. 1992, ALAZARD étal. 1992) and H-NS (FALCONI étal. 1991) histone-like

proteins. Binding of transposase is transient and the influence of histone-like proteins on

this binding has not been directiy investigated.

The three dimentional structure of the transposase N-terminal domain was recently

solved and the lAS binding domain consiste of a novel class of winged helix-turn-helix motif

(CLUBB et al., 1994). In view of the high similarity between repressor and transposase in

that région and the location of ail 4 types of thermosensitive DNA binding repressor

mutations in that région. Mu repressor is very likely to display the same structural

features (see Figure 1).

95

In order to better understand repressor-operator interactions and the features that

distinguish repressor and transposase binding to that région, it is important to isolate and

characterize operator mutations which modify these interactions. The classical in vivo

approach, i.e. the isolation of virulent mutant Mu phages able to overcome Mu immunity and

multipiy in a Mu lysogen, unexpectedly provided very rare repressor mutations but no

operator mutations (GEUSKENS et al. 1991, 1992). One possible explanation for that

observation is that mutations in the operator that prevent repressor binding (O^) aiso

simultaneousiy affect transposase binding and hence block phage DNA réplication and the

ability to form plaques.

Therefore, in order to hâve a first évaluation of the relative importance of the

different operator subsites and their nucléotides for interaction with repressor, we set up a

different in vivo sélection to isolate mutations which allow for constitutive expression from

the early pE promoter. The characterization of these mutations confirmed their location in

the operator région, suggests the existence of a hierarchy in the repressor binding sites

located in the

01

and

02

operators and sheds more light on the relative importance of some

MATERIALS AND METHODS

Bacterial strains. Ail the bacterial strains used in this study were derived from

MC4100 (CASADABAN 1976) which was lysogenized with Mucfs62 (HOWE 1973) or Muc"^

(TAYLOR 1963). MC4100/j/mA;:Tn10 (GAMA étal. 1992) was constructed by transducing

the D82himA::Tn10 mutation from MC253 (M. Chandler's collection) into MC4100 with

phage PI. The pJV300, 304 and 313 plasmids hâve been described (VOGEL étal. 1991,

GEUSKENS et al. 1992). They carry the left end of Mu, with the c"^, cts62 and

cts62,sts62-1 mutations respectively, up to the pE promoter, which is fused to the lac operon (see

results for more details).

Media. Bacteria were grown in Luria-Bertani broth (LB) (Miller, 1972) and titrated on

LB agar plates (MILLER, 1972) containing LB supplemented with 1.2% Difco agar. Minimal

medium was 132 (GLANSDORFF étal., 1965) supplemented with 0.2% lactose. Kanamycin

(Kn, 25 pg ml‘^) and ampicillin (Ap, 25 pg ml*"') were included when appropriate. Mac

Conkey-lactose was from Difco.

»

Chemicals and enzymes. Enzymes were purchased from Boehringer-Mannheim and used

as recommended by the suppliers. ‘

Methods. General methods used in the manipulatin of phages were those of BUKHARI and

LJUNDQUIST (1977).

p-galactosidase was assayed as described by MILLER (1972).

*

Small-scale extraction of plasmid DNA and transformation of that DNA into competent cells

were performed as described by MANIATIS étal. (1982).

Repressor in crude bacterial extracts was assayed as described by GEUSKENS et al. (1991)

A 981

976

I

1004 1011

02

Hind m

GGAATTTACCAAAAAGCAGCmACAAAAAaCTTTTCA&AÆTArCTTTTTAGTAAGC

il II

G (7.4) A(L5) A T (4.1, 4.2, 6.1,8.4, L3)

(1.5, L8)

T) 880 01

repressor translation initiation

ACACCAi^ ATTGACTTTTCAGTATTATTCTTTTCTATAAAGTTACTTTTCAAAAT|TTlAAAC'rCC’r

111111 M 1111 _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

ACACCAP. ATTGAC^

IHF 02

1005

TATTTATCA ÆGCGTTAATCAGTAA rCAAAGGAATTTACCAAAAAGCAGC rTTAc; AA I^AGCTT

M

TT

deleted région ^

rTCAGTAAltatct: 'TTTAGTAAGCTAGCTA\gtttttacacttagttaaattgctaactttat

IMM IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIMIIIIIIIIIIIIIIII

TATCT'; 'TTTAGTAAGCTAGCTA\GTTTTTACACTTAGTTAAATTGCTAACTTTAT

03

Hind III site

AGATT^CAAAACTTAGGAGGGTTTTTAAATGTGTTCCAACGAAAAGGCCCGTGATTGGCATCGT

IIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIIMIIMIIIIMIIIMIIIIIIII

AGATTACAAAACTTAGGAGGGTTTTTAAATGTGTTCCAACGAAAAGGCCCGTGATTGGCATCGT

-35

A. Point mutations. The coordinates are calculated from the first nucléotide at the Mu left

(or c) end (PRESS et al. 1987). The Hind\\\ restriction site is marked by a shaded box. The

horizontal arrows delineate the consensus repressor binding sequences. Among the mutants

analyzed eonly 4.1 and 4.2 originated from the same mutagenized culture and could be

siblings.

B. Délétion mutation. The 01, 02, 03 and IHF binding sites are boxed (open boxes for

operators, hatched box for IHF) according to footprinitng data published earlier (Krause and

Higgins 1986, Alazard et al. 1992). Coordinates are as in A. The délétion in mutant 1.3 is

shown by the thick horizontal arrow. The pE and pCM promoters are shown with their -35

and -10 postitions. The new promoter generated after the délétion could hâve its -35

position at the border of the délétion (TTGACT starting at position 887) and its -10 in the

1022-1032 région.

C. The plasmid used for selecting operator mutations. The left end of Mu up to the pCM

promoter was inserted in pRS551, a pBR322 dérivative (SIMONS étal. 1987), at the EcoRI

site (VOGEL et al. 1991). In this construct the Mu c gene is expressed from the pCM prmoter

97

RESULTS

Sélection of mutations allowing for constitutive expression from pE.

Using plasmid pRS551 (a pBR322 dérivative, SIMONS et al 1987), VOGEL et al.

(1991) constructed plasmids which contain the left end of Mu, including the repressor gene

and the whole regulatory région. In these constructs, repressor is expressed from its

naturel promoter pCM and p_E is fused to the lac operon (see Figure 2C). In bacteria which

hâve a chromosomal délétion of the lac operon, the plasmids should allow for sélection of

cis-dominant constitutive expression from pE, by isolating bacteria able to grow with

lactose as the sole carbon source.

The pRS551 dérivatives pJV300, pJV304 and pJV313 were introduced into MC4100.

The three plasmids respectively carry the wiid type, cts62 and cfs62,sfs62-1 repressor

genes. The cts62 mutation is a thermosensitive mutation which confers thermoinducibility

to Mucfs62 prophages and to pE-initiated lacZ expression from pJV304. The sfs mutation

(for suppression of température sensitivity) is an amber mutation which truncate the

repressor C-terminal end by 18 amino acids. Prophages with a cfs62,sfs62-1 mutated

repressor gene are no more induced at high température. Moreover, contrary to the cts62

repressor which binds Mu operator DNA very poorly at 42°C, the purified cfs62,sfs62-1

repressor binds operator DNA at 42°C almost as efficiently as wild type repressor (VOGEL

et al., submitted for publication). Although there is no clear understanding of the molecular

basis of the Sts phenotype yet, ail the available experimental résulta suggest that it behaves

as a "super-repressor" (VOGEL et al., submitted for publication). Ail these strains grew

well on minimal lactose medium at ail températures. This is most likely due to the fact that,

on the plasmids used, pE is not fully repressed (VOGEL et a/.,1991). Since there are

indications that the host Clp protease destabilizes Mu repressor (GEUSKENS et al. 1992, our

unpublished résulta, SHAPIRO 1993), we transformed the three plasmids in a

MC4100c/pP::Cm strain. In that genetic background, repression of lacZ expression on

pJV313 (with the cts.sts c gene) was indeed strong enough to prevent growth on minimal

lactose medium.

Several independent cultures of MC4100c/pP::Cm/pJV313 were grown in LB,

mutagenized with ethyl-methane-sulfonate (EMS, which cause primarily AT->GC

transitions) and plated on minimal lactose plates which were incubated at 30°C. Control

plates of the same cultures not treated with EMS revealed that the overall mutation

frequency was high (around 10'^) and the same in ail cultures. Consequently, it is not clear

whether the mutants recovered arose from induced or spontaneous mutations.

Isolated colonies were purified on MacConkey lactose medium at 30°C; ail colonies

were red. Then 63 mutant plasmids recovered from 12 independent mutagenised cultures

were transformed into MC4100(Mucfs62) in order to check for the cis-dominance of the

Each Slot has been loaded with 5 pi of a bacterial crude extract.The gel is

12

% on acrylamide.

l.MC4100/pJV313(pRS cts62,sts62-1 ); 2.MC4100 c/pP::CM/pJV313;

3.MC4100 c/pF::CM/pRSL3; 4.MC4100 c/pP::CM/pRSL5 and

98

mutations. Ail the transformée! colonies were still red on MacConkey lactose plates at 30°C

while the MC4100(Mucfs62)/pJV313 was white to pink.

Physical characterisation of the mutations.

Plasmid DNA was extracted from the 63 mutant strains and digested with Hind\\\ which

cuts at a site which overlaps the second and third repressor binding subsites in

02

and hence

could hâve been altered in a subset of the operator mutations. Indeed 55 plasmids had lost

that site, while the other seven remaining ones had the same restriction pattern as the

parental pJV313. These seven plasmids and six which had lost the Hindlll site were further

analysed by DNA sequencing. This confirmed that ail c genes still carried the cts62 and

sfs62-1 mutations. Changes in the operators are shown in Figure 2. 13 mutants, 12 of

which came from independent cultures, were sequenced. Nine are in the 02 operator. Eight of

them are C->A transversions at three sites in two consensus repressor binding sites (the

first and third in 02) and one is an A->G transition at the very beginning of 02 (as defined

by foot-printing experiments, KRAUSE and HIGGiNS 1986, ALAZARD étal. 1992). The four

délétion mutants sequenced, which originated from independent cultures, carry the same

délétion of 01, the UHF binding site and the first subsite of 02. This délétion probably

occured by recombination between two 10-bp direct repeats (CTTTTCAGTA) which

constitute the first 01 and third 02 subsites. This leaves the latter unaffected but should

delete the -35 région of the pE promoter and the repressor translation initiation codon (TTG

see Figure 2B). Consistent with the fact that this mutant expresses |3-galactosidase there is a

potential new promoter with a -35 box (TTGACT) at position 887 and a -10 position around

position 1025.

Expression from pE and pCM on the mutant plasmids.

The ability of the different mutant plasmids to synthesize repressor was tested. As

shown in Figure 3, western blot analysis of repressor présent in the different mutant

strains showed that none of the point mutations prevented repressor synthesis. Whether

expression from pCM and pE occured in the same cell and, in that case, from the same or

different copies of the plasmid remains an open question.

pE driven p-galactosidase expression was measured at 30 and 42°C in exponentially

growing cultures of MC4100 strains carrying one mutated plasmid of each class (pRS1.3,

7.4, L3, L5 and L

8

). As shown in Table 1, pRSL5 has the highest level of constitutive

expression from pE at 30°C, followed by pRSL3, L

8

and 7.4. Since in pRS1.3 the pE

promoter is disrupted, it is not surprising that lacZ expression is not significantly affected

by température. In the other mutant plasmids, the pE -35 and -10 régions remained

unchanged. The parental plasmid pJV313 is poorly induced at 42°C due to the presence of the

sfs62-1 mutation which suppresses the cts62 thermosensitivity. In contrast, ail the mutant

sfs62-1 mutation. Similar values were obtained in strains which carried a Muc+ or a

Mucfs62 chromosomal prophage (data not shown) which is not surprising since, as shown

above, ail the mutated plasmids but pRS1.3 do express their own repressor gene.

p-galactosidase Miller units

30°C 42°C induction

factor

constitutive

expression

MC4100c/pP/pJV313 17 851 50

-MC4100c/pP/pJV304 99 2323 23

-MC4100c/pP/pRS7.4 262 7226 27 15

MC4100c/pP/pRSL8 1 070 6673

6

63

MC4100c/pP/pRSL3

1 1

80 9000 9 69

MC4100c/pP/pRSL5 2184 1 4624 6.7 128

MC4100c/pP/pRS1.3 1 690 2937 1.7

-Table 1.

Overnight cultures grown in LB (Miller 1972) at 30°C were diluted 1:100 in fresh LB and

grown with aération at the same température to early exponentiel phase (

210

®

bacteria/ml). The cultures were split in two halfs. One was kept at 30°C, the other was

shifted to 42°C. p-galactosidase, expressed in Miller units, was assayed according to Miller

(1972) before the splitting of the cultures and 15 or 30 min. after the shift in both the 30

and 42°C cultures. Bach number is the mean value of at least two measurements. The last

column gives the ratio of p-galactosidase production by the mutants compared to the parental

plasmids at the non inducing température (30°C).

100

DISCUSSION

AN the point mutations selected by constitutive expression from pE map in 02 and ali

except one lie at three positions in what was earlier suggested to be the consensus repressor

binding site. Mutations at these three positions are CG to AT transversions in the first and

the third 02 subsites suggesting an important rôle for these three G nucléotides in

repressor/DNA contacts. Surprisingly the G's mutated in L5 and L3 (and other indépendant

isolâtes of the same mutation) are in the same nucléotide of the undefined portion of the

Cl I I INNNWWW consensus sequence. This suggests that there may be some hierarchy in the

subsites and that a G at the third N position might mark subsites which are more important

in initiating or maintaining the repressor/operator complex. Consistent with this

hypothsis, footprinting experiments performed in other laboratories suggest that the third

02 subsite is the most protected from méthylation by DMS (ZHU and HIGGINS, unpublished

results) and the first to be occupied at low repressor concentration (ROUSSEAU et al., to be

published). Moreover, ROUSSEAU et al. (idem) aiso found that the first repressor contact in

01 occurs at the first consensus binding sequence in that operator which aIso carries a G at

that position. However, because mutations of the L

8

and L3 types lie in between the -35 and

the -10 régions of the pE promoter we can not exclude for these two classes that RNA-

polymerase binding to pE is aiso influenced by the mutations. Only carefull analysis of the

affinities of both RNA-polymerase and repressor to the mutated operators will definitely

clarify that point.

It is interesting that the first structural analysis of the complex formed between the

lAS binding domain of Mu transposase made use of the 5'-dTAGCTTTTTAGTAA-3'

oligonucleotide which contains a G at that particular position (CLUBB et al., 1994). It would

be of interest to look at the effect of a transition replacement at that position on

transposase/IAS interaction.

Mutation 7.4 changes an AT to a GC base pair at the left end of 02. This portion of the

early regulatory région is located between 01 and 02, contains the IMF binding site, and is

known to be naturally curved (HIGGINS et al. 1989, BÉTERMIER et al., in press). It is

further bent upon either IHF (HIGGINS étal. 1989, ALAZARD étal. 1992) or repressor

binding (BÉTERMIER et al., in press) or binding of both proteins. A change from A to G in

Documentos relacionados