Involvement of Phage T5 Tail Proteins and Contact Sites between

T H E JOURNAL OF BIOLOGICAL CHEMISTRY 0 1992 by The American Society for Biochemistry and Molecular Biology, Inc.

Vol. 267, No. 5, Issue of February 15, pp. 3173-3178, 1992 Printed in U.S.A.

Involvement of Phage T5 Tail Proteins and Contact Sites between the Outer and Inner Membrane of Escherichia coli in Phage T5 DNA Injection*

(Received for publication, April 17, 1991)

Gilles Guihard, Pascale Boulanger, and Lucienne LetellierS From the Laboratoire des Bwmembranes, UA 1116 Centre National de la Recherche Scientifique, Universite Paris Sud, Bat 433, 91405 Orsay Ceder, France

The penetration of phage T5 DNA into the Esche- richia coli envelope takes place through ion channels (Boulanger, P., and Letellier, L. (1992) J. Biol. Chem. 267, 3168-3172). To identify putative phage protein(s) involved in the formation of these channels, E. coli cells were infected at 37 OC with radioactively labeled phage and their envelopes were fractionated. After a flotation gradient, proteins belonging to the phage tail were recovered both in fractions containing the contact sites between the inner and outer membranes and in the outer membrane. The electrophoretic banding pattern of phage proteins indicates that the contact sites were enriched in the protein pb2. More- over, infected cells were significantly enriched in contact sites as compared to intact cells. There was no enrichment of contact sites and very little radioactivity was found in this fraction and in the outer membrane when the cells were infected at 4 O C (i.e. under conditions where the phage does not inject its DNA). These results suggest that both contact sites and pb2 may play a central role in the translocation of phage T5 DNA.

Phage T5 belongs to the family of the T-odd phages. It possesses a flexible noncontractile and nonpenetrating tail (Bradley, 1967). Its attachment to the Escherichia coli outer membrane receptor (the F h d protein) (Braun et al., 1973) which is mediated by a minor tail protein (pb5) (Heller and Schwarz, 1985) is followed by the transfer of the DNA through the envelope. In the preceding paper (Boulanger and Letellier, 1992) we present experiments which suggest that the DNA of phage T5 may be injected into the cytoplasm through channels in the host cell envelope. We previously proposed that channels are also involved in the penetration of phage T4 DNA (Boulanger and Letellier, 1988; Letellier and Boulanger, 1989). However, the differences in the structural characteristics of T5 and T4 channels make it unlikely that a single protein in the E. coli envelope is responsible for the translocation of the DNA of both phages. We favor the hypothesis that at least some of the channel components are carried by the phage. This paper is devoted to the identification of these putative proteins and to their localization in the envelope.

* This work was supported by the Centre National de la Recherche Scientifique, UA 1116, and by the European Economic Community under the program “Science” (Contract SC10334-C). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “adver- tisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

4 To whom correspondence should be addressed.

Phage T5 injects its double-stranded DNA in two steps (Lanni, 1968; McCorquodale and Warner, 1988): first, there is the entry of the first step transfer (FST)’ DNA which represents 8% of the total T5 chromosome; then, after a pause of a few minutes during which proteins coded by this fragment are synthesized, the remaining 92% of the DNA or second step transfer (SST) DNA is injected. The injection of DNA stops at the FST stage if the infection occurs in a buffer without an energy source or in the presence of chloramphenicol; under these conditions, the DNA remains in continuity between the cytoplasm and the capsid and it is possible to remove the phage capsid by successive centrifugations or by blender treatment of the phage-bacterium complexes (Labe- dan, 1976; Labedan et al., 1973). When the capsid is removed under gentle conditions (i.e. by centrifugations) the uncoiled double-stranded post-FST DNA remains in the external medium (Labedan et al., 1973). This DNA fragment can be translocated across the envelope if the phage-bacterium complex is transferred into a medium allowing protein synthesis (Labedan et al., 1973) indicating that the mechanism of DNA translocation is not significantly impaired by such treatment. These characteristics make phage T5 infection a good experimental model for investigations of the putative channel in the envelope since most of the phage proteins ( i e . the major tail and head proteins pb6 and pb8 (Zweig and Cummings, 1973)) which are not believed to be involved in DNA translocation can be removed by blending, without removing those proteins which may be required for DNA transfer.

E. coli cells were infected at 37 “C with phages containing radioactively labeled proteins; after injection of the FST DNA fragment, capsid of the phages were removed, and the cells were broken in a French pressure cell and fractionated according to the protocol described by Ishidate et al., 1986. This method not only permits the separation of the inner membrane from the outer membrane but also allows the isolation of the so-called “outer membrane light fraction.” This fraction which contains peptides of the inner and outer membranes is likely to correspond to the membrane adhesion sites seen in electron micrographs of plasmolyzed cells (Bayer, 1979; Ishi- date et al., 1986). Ishidate et al. (1986) have given several arguments which suggest that this fraction represents a native structure and does not arise from secondary reassociation of the inner membrane with the outer membrane-murein layers of the cell envelope. The experiments described in this paper show that a phage protein is specifically recovered in the contact sites between the inner and outer membrane. Fur-

l The abbreviations used are: FST, first step transfer; SST, second step transfer; HEPES, N-2-hydroxethylpiperazine-N’-2-ethanesul- fonic acid; OM,, outer membrane, light fraction; OMh outer membrane, heavy fraction; IM, inner membrane.

3173

3174 Phage Proteins Interact with Contact Sites during DNA Transfer

thermore, the amount of contact sites recovered from infected cells is higher than from uninfected cells. The roles of the phage tail protein pb2 and of the contact sites in the translocation of phage DNA are discussed.

EXPERIMENTAL PROCEDURES

Preparation of Radioactive Phage Stocks and E. coli Cells-At- tempts to iodinate a phage preparation were disappointing: the phage proteins were highly iodinated but more than 80% of the population had lost its ability to infect the bacteria. Therefore phage were labeled with ["S]methionine.

E. coli F (Lanni, 1960) cultures were grown in 2.3 BT (Boulanger and Letellier, 1992) to an A m of 0.5 (5 X 10' cells per ml). They were then centrifuged (10 min, 6000 X g), and the pellet was washed once in M9 minimal synthetic medium (Boulanger and Letellier, 1992), pH 7, and resuspended to an As0 of 1.5 in the same medium supplemented with 0.4% glucose, 100 pglliter thiamine and a mixture of amino acids. Each amino acid was added to a concentration of 500 p M except cysteine and tyrosine (20 pM) and methionine (5 pM). The cells were incubated at 37 "C for 15 min and then infected at a multiplicity of 5 plaque forming units per bacteria. 10 p M of [3sS] methionine (6.3 MBq/mmol) was added 12 min later (i.e. a t time when the first structural proteins of phage T5 appear) (Zweig and Cummings, 1973). Phage particles were collected after lysis and purified as previously described (Yamamoto et al., 1970). Phage stocks were stored in "phage buffer" (10 mM Tris, pH 7.4, 100 mM NaCl, 1 mM MgS04, 1 mM CaC12 at a concentration of about l O I 3 infectious particles per ml. The specific activity of the phage stocks was -0.3 MBq/lO'* phages.

The protocol for the preparation of phages with label DNA (La- bedan and Legault-Demare, 1974) was as above except that the medium was supplemented with casamino acids and that [3H]uridine (1 TBq/mmol, 30 nM) was added at the time of infection.

For the preparation of radioactive E. coli F, the cells were grown in minimal medium and in the presence of [3H]leucine (0.6 p ~ , 1.26 TBq/mmol) for three generations.

Preparation of "FST Complexes""Cel1s were grown in M9 medium supplemented with 0.4% glucose to a density of 5 X 10' cells per ml. They were then centrifuged, washed once, concentrated 10-fold in M9 and incubated at 37 "C for 10 min without agitation in the presence of chloramphenicol (50 pg/ml final concentration). Radio- active T5 phage was then added at a multiplicity of infection of 20. Under these conditions 99% of the phage were adsorbed but only the FST DNA was injected (Lanni, 1968). 5 min later, the FST complexes were disrupted in a Sorvall Omni mixer by three successive blendings (each of 2 min at 4 "C) separated by 1 min of incubation at 4 "C. The complexes were then centrifuged (6000 X g, 10 min, 4 ") and washed twice in M9 containing DNase (10 pg/ml) and 10 mM MgSO,. The pellets were resuspended at a density of 4 X 10" cells/ml in 10 mM HEPES, pH 7.4, containing 20% (w/v) sucrose, DNase, and RNase (each 10 pg/ml).

The protocol for the preparation of FST complexes at 4 "C was as described above except that the phage was allowed to adsorb for 15 min at 4 "C before blending.

Fractionation of Membrane Domains-FST complexes were fractionated and purified at 4 "C according to Ishidate et al. (1986) with slight modifications. The FST complexes were broken twice in a French pressure cell (10,000 Ib/in2); the unbroken cells were removed by centrifugation and the supernatant was loaded on a two-step sucrose gradient (SGO) and centrifuged for 35 min. in a TLlOO Beckman centrifuge (rotor SN637, 541,000 X g). The material collected which contained about 2.5% of the initial radioactivity was separated on an SG1 gradient as previously described (Ishidate et al., 1986). The fractions were collected from the top using an LKB Auto Densi-Flow apparatus. Aliquots were taken for the determination of the radioactivity (phage proteins) and for the measurement of the tryptophan fluorescence emission (L,,, = 280 nm; L,,, = 320 nm) (total protein content; in these experiments the amount of phage protein represented only 0.01% of the bacterial proteins). A control experiment performed with [3H]leucine-labeled E. coli membrane vesicles showed that the distribution of the radioactivity overlapped that of the tryptophan fluorescence (not shown). The fractions were pooled according to their location on the gradient, applied to the bottom of a sucrose gradient, and subjected to a density flotation gradient (SG2) (Ishidate et al., 1986). Fractions were collected from the top and concentrated by centrifugation (TL100, 4 "C, 30 min,

541,000 X g) after addition of MgSO, (6 mM) in order to chelate the EDTA present in the buffer. Such treatment favors the aggregation of membranes and therefore increases the amount of sedimented material. The protein content was determined as described elsewhere (Kaplan and Pedersen, 1989).

Disruption of Phage Particles-Phage particles were disrupted by repeatedly freezing the suspension in liquid nitrogen and thawing in a 46 "C water bath (Zweig and Cummings, 1973). The broken phage preparation was then incubated for 1 h at 37 'C in the presence of DNase (10 pglml) and 10 mM MgSOI.

Polyacrylamide Gel Electrophoresis and Autoradiography-Electro- phoresis was carried out essentially as described by Laemmli (1970); resolving mini gels containing 8.5% acrylamide were used. The gels were dried and autoradiography was performed by exposing the gel to a hyperfilm Bmar (Amersham) for 1 to 2 weeks. Each phage protein band was identified according to Zweig and Cummings (1973).

RESULTS

Fractionation of Phage-Bacteria Complexes-E. coli cells were infected at 37 "C with [3sS]methionine labeled phage under conditions allowing only the injection of the FST DNA (Lanni, 1968), and the phage-bacteria complexes were disrupted by blending. Only 10-15% of the radioactivity re- mained associated with the complexes and there was no significant lysis as judged from the absorbance of the suspension. The cells were blended either at 4 "C or at 25 "C. Since the distribution of the membrane proteins and of the radioactivity on the SG1 gradient (see below) was similar at both temperatures, the blending was subsequently done at 4 "C for all conditions of infection. After blending, the infected cells were centrifuged and the proteins recovered in the supernatant analyzed by gel electrophoresis. The autoradiogram of the labeled phage proteins after sodium dodecyl sulfate-polyacrylamide gel electrophoresis is presented on Fig. 1: the bands corresponding to the tail proteins pbl (the protein forming the L-shaped fiber and the ring-like structure around the tail and which serves to the reversible attachment to the outer membrane) (Heller and Braun, 1979; Saigo, 1978), pb2 (the straight fiber) (Saigo, 1978), and pb5 (involved in receptor binding) (Heller and Schwarz, 1985) were not detectable; the intensity of the bands corresponding to pb3 and pb4 whose function is not clearly defined (Feucht et al., 1989) were much lower than in the intact phage. The banding pattern of sam- ples in which the cells were infected at 4 "C (i.e. under conditions allowing only the irreversible attachment of the phage to the bacteria) were similar to those in which infection

37' 4' T5

200 m ds.,

/" -2 -3

4 8

a b c d FIG. 1. Autoradiogram of a n 8.5% polyacrylamide gel con-

taining T5 proteins labeled with 35S. a, standard proteins; b and c, phage proteins released into the supernatant after centrifugation of the phage-bacteria complexes disrupted at the FST stage respectively at 37 and 4 "C. d, proteins obtained from phage particles disrupted by freezing and thawing. The numbers on the right refer to the T5 polypeptides pbl to pb8 (see text). The numbers on the left refer to the molecular mass (in kDa) of standard proteins.

Phage Proteins Interact with Contact Sites during DNA Transfer 3175

was at 37 "C. This suggests that these tail proteins remain attached to the bacteria even under conditions where the DNA is not injected.

The phage-bacteria complexes were fractionated as described under "Experimental Procedures." The radioactive and protein profiles obtained after the SGl sedimentation gradient are represented in Fig. 2A (infection at 37 "C) and 2B (infection at 4 "C); as a comparison, the SG1 profile of uninfected E. coli F cells is represented on Fig. 2C. The three peaks identified by Ishidate et al. (1986) as the inner membrane (IM), the outer membrane light fraction (or contact sites) (OM1) and outer membrane heavy fraction (OMh) were clearly recovered. Their similarity with the fractions obtained by Ishidate et al. (1986) was infered from the sodium dodecyl sulfate-gel electrophoresis pattern of the membrane proteins, the determination of the apparent buoyant densities, the lipopolysaccharide content and the respiratory activity (data not shown). Interestingly, the amount of membrane protein recovered in the contact sites after infection at 37 "C was at least twice that recovered after infection at 4 "C (result of six independent experiments). Radioactivity was present in all membrane fractions from infection at 37 "C but little radioactivity was recovered in the OM1 and OMh fractions from infection at 4°C. The peak of radioactivity found in the inner membrane did not coincide with the membrane protein peak (result of six fractionations) suggesting that it may have originated from contaminating material which may have failed to reach equilibrium during the centrifugation (see below). A fourth peak (the "phage" peak) which contained few bacterial proteins but a significant amount of radioactive material was found between peaks IM and OM]. This peak was phage material which comigrated with the membranes in the gradients since it was recovered at the same apparent buoyant density when phage proteins either alone or mixed

- t n

0 10 20 30 fraction number

- r uninfected cells OM.

fraction numbel

FIG. 2. Radioactive and protein profiles after the SG1 sedimentation gradients of cells infected at 37 OC ( A ) , 4 "C ( B ) , and uninfected (C). The radioactivity refers to labeled phage proteins. The protein profile (membrane proteins) was deduced from the tryptophan emission fluorescence (L,,, = 280 nm; X,, = 320 nm). IM, phage, OW, and OM,,, correspond, respectively, to the inner membrane, phage peak, outer membrane light, and heavy fraction.

with French-pressed E. coli vesicles were layered on a SG1 gradient (Fig. 3) .

All the fractions of each peak from the SG1 gradient of infection at 37 "C were independently pooled and further purified on SG2 density flotation gradients (Ishidate et al., 1986). Fig. 4 shows that each initial fraction could be resolved into several fractions. The "phage" peak was resolved into two peaks, one sedimenting at the apparent buoyant density of the inner membrane which contained no radioactivity and very few membrane proteins; a second peak (d = 1.250) contained almost all the radioactivity (Fig. 4A). Fig. 4B shows that all the radioactivity which was associated with the inner membrane (infection at 37 "C) after the SG1 gradient mi- grated at a density corresponding to the phage proteins (d = 1.250) confirming that it originates from contaminating phage particles. A similar pattern was obtained with the inner membrane isolated from the cells infected at 4 "C (not shown). The OM1 fraction was resolved into three peaks corresponding respectively to the OMI, OMh, and phage proteins (Fig. 4 c ) . Fraction OMh was resolved into one major peak corresponding to the OMh and a small shoulder which probably originates from OM1 (Fig. 4 0 ) . Interestingly, the specific radioactivity of the OM1 was at least twice that of the OMh fraction (result of four different fractionations). The OM1 and OMb fractions from the cells infected at 4 "C were not purified on SG2 because insufficient material was recovered from the SG1 gradient.

Fig. 5 shows the autoradiogram of the proteins recovered in the OMl and OMh fractions from infection at 37" and 4 "c. The relative intensities of the polypeptide bands of the phage tail (pbl to pb5) were comparable in the control and in the OMI and OMh originating from the SG1 fractionation of infection at 4 "C. The OM1 and OMh fractions from the SG2 gradient of infection at 37 "C were significantly enriched in pb2. This enrichment was much more pronounced in the OM1 than in the OMh fraction; a significant enrichment in pb4 was also observed in the OM1 fraction.

The presence of phage DNA associated with the membrane fractions was demonstrated by infecting cells at 37 "C with T5 phage containing labeled DNA. After fractionation on a SG1 gradient, a first peak of radioactivity was found between fractions 1 and 10 which only partially overlapped the inner membrane peak (Fig. 6); furthermore after centrifugation of these fractions, the radioactivity was only recovered in the supernatant suggesting that it originates from free oligonucle- otides and was not associated with the inner membrane. In contrast, all the radioactivity which overlapped both the OM1 and OMh fractions could be sedimented which indicates that it was associated with these membrane fractions. Further- more, the specific radioactivity of the OM1 was 1.5 times

fraclion number

FIG. 3. Radioactive profiles of phage TS proteins. Phage proteins were dissociated by several cycles of freezing and thawing. They were either mixed with French-pressed E. coli vesicles (FPV), purified first on a SGo gradient and then on a SG1 gradient (T5 + FPV, .) or directly purified on a SG, gradient in the absence of membrane vesicles (T5,O).

3176 Phage Proteins Interact with Contact Sites during DNA Transfer

fraction number

fraction number

1.2!5

450

300

150

0 0 10

fraction number ao 30

- 600 - 300 E, 3. 0 400 200 ; 2 200 100 5

U .- h

> .- 24

0 .- 0 U

L

e o 0 10 zn 30 4u

o = fraction number

FIG. 4. Radioactive and protein profiles of the SG2 flotation gradient. The fractions from the IM, phage peak, OM,, and OMh fraction after SG1 centrifugation (infection at 37 "C) were pooled and applied to the bottom of sucrose gradients. The protein profile was determined from the tryptophan fluorescence (X,,, = 280 nm; L,,, = 320 nm). The fluorescence intensity of each fraction was measured in comparison to the same fluorescence standard. The numbers over each peak refer to the buoyant density of the fraction.

higher than that of the OMh fraction. No radioactivity was found in the OM1 and OMh fractions when the cells were infected at 4 "C (data not shown).

CONCLUSIONS AND DISCUSSION

Phage T5 contains at least 15 different structural polypeptides (Zweig and Cummings, 1973): the major head and tail proteins pb8 and pb6 represent, respectively, 65 and 17% of the total protein content of the phage. These proteins are not required for the transfer of DNA through the envelope since the infection can take place even after the capsid of the phage has been removed (Labedan et aL, 1973). Polypeptides pbl to pb5 are in the phage tail: each of these tail proteins represents 1 to 2% of phage protein (McCorquodale and Warner, 1988;

370 - - 40

1 5 OMh OM1 Oh+, OM1

:= 3- 4-

6- ' 5"

FIG. 5. Autoradiogram of the banding pattern of phage proteins associated from the fractions obtained after the SGI sedimentation gradient (OMI and OMh at 4 "C) and SG2 gradient (OM1 and OMh at 37 "C). Aliquots of the different fractions were loaded on an 8.5% polyacrylamide gel: each fraction contained 60 pg of proteins and 7000 f 1000 cpm except the OMh sample at 4 "C which contained 1000 cpm. The first banding pattern corre- sponds to phage T5 standard proteins. The migration of the bands of the OM1 and OMh fractions is slightly disturbed due to the presence of lipopolysaccharide.

fraction number

FIG. 6. Radioactive and protein profiles of phage-bacteria complexes containing radioactively labeled phage DNA on a SGI sedimentation gradient. All other conditions are similar to those described in the legend to Fig. 2.

Zweig and Cummings, 1973); since the bacteria can be infected with a limited number of phages only, this makes the detection of the putative phage protein(s) in the envelope rather diffi- cult. Some of the difficulties were overcome: first, using [3sS] methionine labeling, phage T5 particles with high specific activity were produced; second, most of the major tail and head proteins were eliminated by blending after injection of the FST DNA which removed 85-90% of the total radioactivity. A critical step in these experiments was to discriminate the phage proteins involved in the reversible and irreversible attachment of the phage from those involved in the translocation of DNA. This was done by comparing the results obtained after fractionation of the cells infected at 37 "C (a temperature at which the DNA is injected) with those at 4 "C (a temperature at which the phage binds to its receptor and the DNA is ejected from the head but remains attached to the outer membrane surface) (Labedan, 1976).

After fractionation of the infected cells on an SG1 sedimentation gradient the different domains identified by Ishidate et al. (1986) as the inner membrane (IM), outer membrane (OMh) and contact sites (OMJ were recovered suggesting that infection did not interfere with fractionation. Interestingly, the amount of membrane protein recovered in the contact site fraction was twice as large in cells infected at 37 "C than in those infected at 4 "C or in uninfected cells. This suggests that infection leads to the formation of new contact sites or to the stabilization of pre-existent ones. These results are consistent with several observations by electron microscopy of plasmolyzed E. coli cells. Lopez and Webster (1985) have presented evidence that the site of assembly and of extrusion of phage F1 occur at sites where the outer membrane and inner membranes are closely joined. Furthermore, they calculated that the number of contact sites per bacteria rose from 350 to 590 upon infection. Similar observations were made in the case of the bacteriophage MS2 (Walderich et al.,

Phage Proteins Interact with Contact Sites during DNA Transfer 3177

1988; Walderich and Holtje, 1989). This phage codes for a hydrophobic protein (the L protein) which triggers the lysis of the host. An increase of 17% in the number of contact sites was observed after expression of the cloned L protein (Wald- erich et al., 1988) and immunoelectron microscopy showed that the L protein is part of the contact sites (Walderich and Holtje, 1989). Recently, Kellenberger (1990) claimed that the contact sites were artifactually induced by the plasmolysis of the cells. The fractionation technique of Ishidate et al. (1986) does not involve any plasmolysis of the cells so that we believe that the presence of the contact sites and the appearance of new contact sites upon infection may reflect physiological processes. This technique, however, requires that the cells would be broken by a passage through a French pressure cell. Although the authors have presented several arguments suggesting that the contact sites represent a native structure, one cannot exclude that the French press treatment causes an artificial reassociation of the inner and outer membranes.

The specific radioactivity recovered in the OM, and 0% fractions of infection at 37 “C was higher than those at 4 “C suggesting that the enrichment is related to the process of DNA injection. Furthermore, for cells infected at 37 “C the specific radioactivity of the OM* fraction was at least twice that of the OMh fraction which suggests that contact sites are specifically involved in the translocation of DNA. The involvement of the contact sites in the translocation of phage DNA is further supported by the presence in this fraction of radioactive phage DNA. The electrophoretic banding pattern of the different fractions indicates that the enrichment in pb2 was large in the OM, fraction and also significant in the OMh fraction when the infection was performed at 37 “C; such enrichment was not found at 4 “C. It is thus tempting to assign a specific role to the protein pb2 in the translocation of phage DNA. The OM1 fraction was also significantly enriched in pb4 whereas its presence was not detectable in cells infected at 4 “C. The function of pb4 remains unclear although Feucht et al. (1989) showed that the attachment of phage T5 to FhuA results in the covalent cross-linking of three copies of pb4. These results suggest that pb2 may remain in close contact with pb4 when the DNA is translocated.

pb2 is the only protein of the straight fiber; this fiber which is probably formed by 5 or 6 copies of pb2 (Feucht et aL, 1990; McCorquodale and Warner, 1988) appears in electron micrographs as a single structure of 50-nm length and 5-nm width consisting of stacked discs (Saigo, 1978). The length of the pb2 multimer is therefore sufficient to span the whole envelope and thus to gain access to the inner membrane. A continuous structure of this type between the outer membrane and inner membrane would explain why the injected DNA remains totally protected from periplasmic nuclease activities (Heppel, 1967).

This model is supported by the recent results of Feucht et al. (1990). They showed that isolated pb2 could form pores when reconstituted into lipid bilayers; from the conductance data, they calculated that the diameter of the pore (assuming that it is a cylinder 7.5 nm long) would be 2 nm; this is large enough to allow the passage of the double-stranded phage DNA.

The experiments described above do not allow the origin of the contact sites involved in the translocation of phage T5 DNA to be determined. The simplest model supposes that every FhuA receptor present on the outer membrane [there are about 400 FhuA molecules per cell (Dubertret and Legault- Demare, 1978)], is “competent” both for the attachment of the phage and the translocation of the DNA. This implies that phage attachment to the receptor is sufficient to trigger

the penetration of pb2 which would remain attached to the outer membrane via pbl, pb4, and pb5 such that it spans the two membranes. After passage of the complexes through the French press, a fraction containing pb2 attached to fragments of the outer and inner membranes would be recovered; in this model the increased yield of the OM1 fraction results from new contact sites created by pb2. A more complex model supposes that the translocation of the phage DNA only occurs at pre-existing contact sites on the envelope, a hypothesis supported by the observations by Bayer (1979) using electron microscopy. If this is the case, the increase of the number of contact sites on infection could be due to the stabilizaiton of these membrane domains by pb2. Some TonB and To1 mu- tants are tolerant to colicins (Braun et d., 1980; Davies and Reeves, 1975; Eick-Helmerich and Braun, 1989). Such bacteria are also resistant to phage infection although the phage can irreversibly bind to these cells. The To1 proteins are involved in the penetration of the DNA of the filamentous phages F1 and Fd (Boeke et al., 1982). Jakes et al. (1988) have shown that a hybrid protein containing the N-terminal two- thirds of G3p (the phage protein involved in the attachment of phage fl to the outer membrane receptor) and the C- terminal domain of colicin E3 translocated via the To1 system and carried the infectivity of colicin E3. Furthermore, TolB which is required for killing by colicin E3 but not for infection by fl was found to be necessary for the killing activity of the hybrid protein. Direct interactions between To1 proteins, outer membrane proteins, and the colicins have also been recently demonstrated.2 Recently, Brewer et al. (1990) have shown that a peptide segment of TonB specifically interacts with the outer membrane protein FhuA. All these results suggest the existence of a complex machinery crossing the envelope which may in part correspond to the proteins recovered in the contact sites between the two membranes. Specific machinery for the translocation of T5 DNA is consistent with the results obtained, in particularly at low temperature. We have measured the time separating the attachment of phage T5 and of colicin A from the formation of their respective pores in the inner membrane (Boulanger and Le- tellier, 1992; Bourdineaud et al., 1990); this lag time was shown to increase strongly with decreasing temperature and the opening of the pores could not be detected below the order- disorder transition temperature of the lipid hydrocarbon chains; we proposed that this lag time not only represents the time necessary for the recognition of the receptor and irreversible attachment but also the time required by the receptor to “find” the trandocation machinery: the receptor presum- ably diffuses laterally in the outer membrane so as to contact the translocation system; this lateral diffusion would be slowed down by a decrease of temperature because of the low apparent lateral diffusion coefficient of the lipolysaccharide (Miilhradt and Menzel, 1974) and be totally prevented when the lipid became ordered. This would explain why pb2 is not found in the contact sites at 4 “C and why the DNA is not injected.

Acknowledgments-We express our gratitude to Drs. A. Ghazi and E. Shechter for helpful discussions. We thank M. Leroux for excellent technical assistance.

REFERENCES Bayer, M. E. (1979) in Bacterial Outer Membranes, Biosynthesis and

Boeke, J. D., Model, P., and Zinder, N. D. (1982) Mol. Gen. Genet. Functions (Inouye, M., ed) pp. 167-202

186,185-192

* Benedetti, H., Lazdumski, C., and Lloubes, R. (1991) EMBO J. 8, 1989-1995.

3178 Phage Proteins Interact with Contact Sites during DNA Transfer Boulanger, P., and Letellier, L. (1988) J. Biol. Chem. 263,9767-9775 Boulanger, P., and Letellier, L. (1992) J. Biol. Chem. 267,3168-3172 Bourdineaud, J. P., Boulanger, P., Lazdunski, C., and Letellier, L.

Bradley, D. E. (1967) Bacteriol. Rev. 31, 262-280 Braun, V., Schaller, K., and Wolff, H. (1973) Biochim. Biophys. Acta

Braun, V., Frenz, J., Hantke, K., and Schaller K. (1980) J. Bucteriol.

Brewer, S., Tolley, M., Trayer, I. P., Barr, G. C. Dorman, C . J.,

Wormald, M. R. (1990) J. Mol. Biol. 216, 883-895 Hannavy, K., Higgins, C. F., Evans, J. S., Levine, B. A., and

Davies, J. K., and Reeves, P. (1975) J. Bacteriol. 123, 96-101 and

Dubertret, C., and Legault-Demare, J. (1978) FEMS Microbiol. Lett.

Eick-Helmerich, K., and Braun, V. (1989) J. Bacteriol. 171, 5117-

Feucht, A., Heinzelmann, G., and Heller, K. J. (1989) FEBS Lett.

Feucht, A., Schmid, A., Benz, R., Schwarz, H., and Heller K. J. (1990)

Heller, K. J., and Braun, V. (1979) J. Bacteriol. 139, 32-38 Heller, K. J., and Schwarz, H. (1985) J. Bucteriol. 162 , 621-625 Heppel, L. A. (1967) Science 156, 1451-1455 Ishidate, K., Creeger, E. S., Zrike, J., Deb, S., Glauner, B., Mac-

(1990) Proc. Natl. Acud. Sci. U. S. A. 87, 1037-1041

323,87-97

142,162-168

102-117

4,63-66

5126

255,435-440

J . Biol. Chem. 265, 18561-18567

Allister, T. J., and Rothfield, L. I. (1986) J . Biol. Chem. 261, 428- 443

Jakes, K. S., Davis, N. G., and Zinder, N. D. (1988) J. Bacteriol. 170, 4231-4238

Kaplan, R. S., and Pedersen, P. L. (1989) Methods Enzyrnol. 172, 393-399

Kellenberger, E. (1990) Mol. Microbiol. 4, 697-705 Labedan, B. (1976) Virology 75, 368-375 Labedan, B., and Legault Demare, J. (1974) J. Virol. 13, 1093-1100 Labedan, B., Crochet, M., Legault-Demare, J., and Stevens, B. J.

Laemmli, U. K. (1970) Nature 227,680-685 Lanni, Y. T. (1960) Virology 10 , 501-513 Lanni, Y. T. (1968) Bacteriol. Rev. 32, 227-242 Letellier, L., and Boulanger, P. (1989) Biochirnie 71, 167-174 Lopez, J., and Webster, R. E. (1985) J. Bucteriol. 163, 1270-1274 McCorquodale, D. L., and Warner, H. R. (1988) in The Bacteriophages

(Calender, R., ed) VoI. 1, pp. 439-475, Plenum Press, New York Miilhradt, P. F., and Menzel, J. (1974) Eur. J. Biochern. 43,533-539 Saigo, K. (1978) Virology 85, 422-433 Walderich, B., Ursinus-Wossner, A., van Duin, J., and Holtje, J. V.

(1988) J. Bacteriol. 170, 5027-5033 Walderich, B., and Holtje, J. V. (1989) J. Bucteriol 171, 3331-3336 Yamamoto, K., Alberts, B., Benzinger, R., Lawhorne, L., and Treiber,

Zweig, M., and Cummings, D. J. (1973) Virology 51,443-453

(1973) J. Mol. Biol. 75 , 213-234

G. (1970) Virology 40, 734-744

Documents

Involvement of Phage T5 Tail Proteins and Contact Sites between