75, 368-375 (1976)
A Very Early Step in the T5 DNA Injection
B. LABEDAN Laboratoire
des Acides Nucle’iques, Institut de Microbiologic, BBtiment 409, 91405 Orsay, France Accepted August
After phage T5 irreversibly adsorbs onto its specific host cell receptor, its DNA is released from the head of the capsid and appears to be immediately attached to the cell envelope. The DNA-attachment step takes place before the sensitization step to deprivation of calcium and before the transfer of the first-step-DNA fragment. Phages having an aberrant injection are deficient for this DNA-attachment step. The DNA attachment itself seems relatively weak, in contrast to the firm link mediated by the injected FSTDNA fragment. In the attached position, the T5 DNA is protected from host periplasmic nucleases even when it has been freed from its capsid by low-speed centrifugation of the infected bacteria. This naked, attached DNA may inject its FST segment after transfer of the centrifuged bacteria from 0 to 37”. Therefore, this DNA-attachment step appears as a transient state before the penetration of the FST-DNA fragment and the subsequent transient block of the chromosome injection. INTRODUCTION
Little information about the successive steps of injection of the T5 phage chromosome into the host cell are available. Recent in vitro studies about the T5 receptor particles, first isolated and characterized by Weidel et al. (1954) showed that a phage adsorbed onto an isolated receptor particle releases its entire chromosome into the external medium (Zarybnicky et al., 1973). On the other hand, Lanni (1968) demonstrated a two-step DNA injection process, the first-step-transfer (FST) DNA segment coding for proteins responsible for the injection of the post-FST DNA. We demonstrated that, after the first step of DNA transfer, the FST fragment was linked to the bacterial membrane and that the postFST DNA, even if it was exposed by centrifugation of the infected bacteria, remained attached to the bacteria on the external side of the envelope (Labedan et al., 1973). Recently, Labedan and Legault-Demare (1974) showed the existence of aberrant T5 1 This paper is part of a doctoral thesis submitted to the University of Paris XI.
phages which, after normal adsorption, released all their DNA outside the bacterium. These phages mimic, with bacteria, the behavior that all the phages exhibit with the isolated receptors. This last fact prompted us to study the deficiency of these phages. During this work we found, as described in this paper, that the phage chromosome ejected from capsids of irreversibly adsorbed normal phages attached to the cell surface and that this attachment step took place prior to the penetration of the FST-DNA fragment inside the host cytoplasm. On the contrary, the aberrant phages did not attach their DNA. MATERIALS
Bacteriological procedures. Bacteriophage and bacterial strains, media and buffers, growth and purification of phage stocks are described in detail elsewhere (Labedan and Legault - Demare, 1974). Preparation and centrifugation of Escherichia coli-T5stO complexes. A fresh culture of E. coli F was diluted into MGM medium and growth was continued until a concentration of 5 x lo8 cells/ml was attained. The bacteria were centrifuged for 368
Copyright 0 19’76by Academic Press, Inc. All rights of reproduction in any form reserved.
STEP IN T5 DNA
15 min at 5000g and resuspended in MGM buffer at a concentration of 5 x log cells/ ml; in some experiments they were centrifuged again twice in this buffer in order to eliminate most of the periplasmic endonuclease I (Labedan and Legault-Demare, 1974). To bacteria incubated either at 37” or kept on ice at 0” for 15 min, 14C-labeled T5 phages were added at an m.o.i. of about 10, and the incubation was continued for 10 min at 37 or 0”. At this time, the T5infected bacteria were centrifuged for 10 min at 6000g in a Sorvall SS-34rotor at 0”. The supernatants (Sl) were carefully decanted, and the pellets (Pl) were resuspended by pipetting into the initial volume of MGM buffer. The radioactivity contained in Sl and Pl was measured on lOO~1 samples deposited onto Whatman 3MM filter discs. In some experiments, the Pl pellets were resuspended without any stirring, by gently covering the pellet with 0.5 ml of buffer and letting it stand for at least 1 hr at 0” in order to obtain a spontaneous resuspension. Then, the Pl-pelleted bacteria, pipetted or spontaneously resuspended, were processed through successive cycles of centrifugation at 6000g and resuspended by pipetting. Lysis of T5-infected bacteria. The bacteria, infected at 0 or 37”, were transformed into spheroplasts: EDTA, pH 7.8 and lysozyme were added to 5 x 10m3M and 1 mg/ ml, respectively, and after 30 min at 0 or 37”, a good spheroplasting was obtained. Then, in some experiments, these spheroplasts were lysed by adding 0.5% Brij-58 for 15 min at either temperature, and the lysates were centrifuged either at low or at high speed on a neutral 20 to 40% sucrose gradient established above a 60% sucrose cushion according to the technique of Knippers and Stratling (1970). In other experiments, the spheroplasts were lysed by osmotic shock and the cell envelope was purified by differential centrifugation to eliminate some unlysed cells (20 min at 3000g) and cytoplasmic constituents (2 hr at 41,000 rpm in a Spinco 50 Ti rotor). Then the purified envelope was eventually freed from the cytoplasmic
membrane by incubation for 10 min in 2% Triton X-100, 10e2M MgCl, (Schnaitman, 1971) and centrifugation in the Spinco 50 Ti rotor; the Triton treatment was repeated, and the pellets of the outer membrane-enriched fraction were analyzed for phage DNA attachment. Radioactivity determinations. These were made as already described by using a toluene/PPO/POPOP scintillating fluid and a Beckman spectrometer (Labedan et al., 1973). RESULTS
T5 DNA Attachment fuged Bacteria
to Low-Speed Centri-
We have already observed that lowspeed centrifugation may transform part of T5-infected bacteria, arrested at the first step of transfer, into bacteria carrying naked uncoiled phage DNA molecules, by breaking the adsorbed phage capsid at points located along the tail (Labedan et al., 1973). In an attempt to determine the first stages of the T5 DNA injection process, it seemed interesting to compare the events occuring in bacteria infected at 0 and at 37”. Indeed, Lanni (1961) showed that the incubation temperature played a major role in the first-step transfer of phage DNA: The lower the temperature below 37”, the slower the rate of transfer, and at 0” there was no transfer at all; however, at all temperatures, the adsorption seemed normally irreversible. Two questions may be raised about the possible effects of centrifugation on bacteria infected at 0”: (1) Are the adsorbed capsids broken away at 0” as they are at 37”? (2) If they are, where does the DNA they contained go? After the first centrifugation of l*C-labeled T5-infected bacteria, the samples incubated at either temperature released the same percentage of radioactivity in supernatant Sl: 10% at 0” and 11% at 37”. We have already described (Labedan and Legault-Demare, 1974) and I have again verified (data not shown) that, whatever the incubation temperature, nearly all the radioactivity contained in each Sl supernatant consisted only of DNA ejected from
adsorbed aberrant phages and degraded by the periplasmic endonuclease I. This first result implied that (1) there are no full heads released into the external medium; (2) the DNA of normal phages remained attached to pelleted bacteria at 0” as well as at 37”, without knowing if this DNA was exposed or still encapsided. To clarify this last point, the Pl pellets, which then contained 100% of the “adsorbed radioactivity,” as defined by Lanai (1961) and by Labedan et al. (1973), were resuspended by pipetting in the original volume, and then immediately recentrifuged. As shown in Fig. 1, part of the phage DNA was released in both 52 supernatants, more at 37 than at 0”. These S2 DNAs were characterized by two criteria: (1) On neutral sucrose gradient, both DNAs gave large double-stranded fragments, as already shown only for the S2 supernatant obtained from bacteria infected at 37” (Labedan et al., 1973) (see also Fig. 2). (2) If the Pl pellets had been pretreated with pancreatic DNase, the same 100
percentages of radioactivity were obtained in each of the two S2 supernatants, but all the DNA they contained was degraded and sedimented at the top of the sucrose gradient (data not shown). From these results it may be concluded that, at either temperature, the low-speed centrifugation broke part of the adsorbed capsids but the DNA they contained remained attached to the infected bacteria in a naked, DNase- and shearing-sensitive state. The aberrant phages (Labedan and Legault-Demare, 1974) appeared to be deficient for this process of DNA attachment to infected bacteria. After several cycles of centrifugation and resuspension by pipetting (necessary to break away all the adsorbed capsids), 99.5% of the radioactivity adsorbed in Pl pellets may be eluted from infected bacteria incubated at 0” (Fig. 1); on the contrary 8% of the label adsorbed in Pl was never elutable from bacteria infected at 37”. Thus, these last results confirmed the preceding observation of Lanni (1961): At 0 there is no penetration of the FST-DNA segment into the T&infected bacteria. I have verified that there is no noticeable loss of infected bacteria in the successive supernatants during repeated centrifugation (see also Labedan et al., 1973). The DNA Attachment
FIG. 1. Effect of resuspension on centrifuged T5infected bacteria. After centrifugation of W-labeled TSinfected bacteria incubated at 0 or 37”, the Pl pellets were resuspended either spontaneously (A and A) or by pipetting (0 and 0). After resuspension the four samples were again centrifuged and resuspended, always by pipetting.
Made at 0” Is Weak
The firmness of the T5 DNA attachment was tested by spontaneous resuspension of the Pl-pelleted bacteria, infected respectively at 0 and at 37”. Figure la shows that, at 37”, 84% of the DNA initially attached to Pl-pelleted bacteria remained associated with gently and slowly resuspended and recentrifuged bacteria (P2 pellet), but only 36% of the Pl-attached DNA resisted resuspension by pipetting or DNase treatment of the Pl pellet. Figure lb shows that, at O”, the respective percentages were 63% for a spontaneous resuspension and 66% for a resuspension by pipetting. Thus the DNA attachment made at 0” appears as labile, at least after a long (1 hr and more) incubation of the T&bacteria complexes. The DNAs, released at 0” from either the pipetted or the pipetted Pl pellets, were
STEP IN T5 DNA
FIG. 2. Analysis of T5 DNA released from resuspended Pl pellets of bacteria infected at 0”. W-labeled phage DNA from S2 supernatants of spontaneously resuspended (0) or pipetted (0) Pl pellets, was analyzed on a neutral sucrose gradient made and centrifuged as described (Labedan et al., 19’731. In a parallel tube, W-labeled phenol-extracted T5 DNA was used as sedimentation marker (arrow). Forty fractions were collected from each tube and their radioactivities were determined.
further analyzed on a neutral sucrose gradient. Figure 2 shows that the nonpipetted sample gave a majority of intact total DNA molecules (including the FST fragment, the noninjection of which was again verified) and several little peaks measuring about 60 and 40% of the total T5 chromosome, thus corresponding to mechanical breakage products (Labedan et al., 1973); on the contrary, the breakage products were more frequent in the pipetted sample, which contained nearly no intact DNA molecule. These different results were obtained even when the periplasmic endonuclease I was not released by repeated washings of the bacteria before T5 adsorption (Labedan and Legault-Demare, 1974). Thus, these results implied that the attached DNA was not released by some nuclease action, and thus was located in a nuclease (essentially endonuclease D-free envelope site. In other words, the DNA of normal phages, ejected from heads of adsorbed capsids and attached to the infected cell, was throughout protected from periplasmic host cell nucleases, contrary to the DNA of aberrant phages described by Labedan and Legault-Demare (1974). Timing
of the DNA Attachment
Lanni (1961) showed that, at 37”, lop3 M
CaCl, was necessary during some of the first stages of DNA injection to obtain a successful T5 infection: She defined a sensitization step to deprivation of calcium ions, which took place after the adsorption and before or at the same time as the first step of DNA transfer. The evidence showed that, in the absence of calcium ions, some phage DNA was released in the external medium of the infected bacteria (Lanni, 1961). Samples of T&infected bacteria at both 0 and 37” were prepared with or without 10e3 M CaCl, added, and with 2 x 10e3 M EDTA added to complex calcium ions. Table 1 shows the results obtained for the six samples: at 37”, without calcium ions or with EDTA added, the elution of some phage DNA, different from the one of aberrant phages DNA, was confirmed, but, at O”, the deprivation of calcium ions or the addition of EDTA gave no release of radioactive material except the aberrant phage DNA. Thus, the T5 DNA attached at 0 was indifferent to the presence of either calcium ions or EDTA, contrary to the post-FST DNA (37” samples) which clearly needed calcium ions to remain in a stable state. In other words, the waiting sites of the attached DNA and of the post-FST DNA differ by, at least, their reactivity in the presence of calcium ions.
B, LABEDAN TABLE
OF DEPRIVATION INFECTED
Incubation medium with 10m3M CaCl, Incubation medium with 2 x lo+ M EDTA Incubation temperature (“C) Sl”
IONS ON T5-
+- + - - + 0 0 0 T
10.5 11 11.5 28 11.5 31
a Radioactivity released from ‘%-labeled T&infected bacteria as percentage of the total input radioactivity at the initiation of the infection.
the DNA Attachment
tually a physical link between the phage capsid-cell envelope receptor complex and the T5 DNA before the first step of DNA transfer occurs. In a later experiment, I attempted to localize the part of the envelope which may contain the DNA attachment site. Spheroplasts, prepared from bacteria infected by T5 phage at o”, were lysed by osmotic shock; from this lysate, I purified the cell envelope by differential centrifugation and then divided it in two parts. One part was the untreated sample; another part was treated with the neutral detergent Triton X-100 (2%) in the presence of magnesium ions, in order to specifically remove the cytoplasmic membrane as clearly described by Schnaitman (1971); then the two samples were again centrifuged. The radioactivity contained in the pellets of the untreated (50,800 cpm/ml) and treated (50,300 cpm/ml) samples was found identical to the label before treatment (52,500); thus, it seems that the removal of the cytoplasmic membrane does not release the attached DNA.
In a first experiment we verified that T5 phage DNA was indeed attached to the bacterial membrane. 14C-labeled T$infected bacteria, prepared as usual at 0 or 37”, were either untreated or transformed into spheroplasts and lysed by addition of the neutral detergent Brij-58. After checking for total lysis by microscope examination, the untreated and lysed bacteria were centrifuged for 15 min at 6000g at o”, and the radioactivity released in the su- Penetration of the Exposed FST-DNA pernatants Sl was determined: The same Is the DNA attachment step an artifact figures were found in the samples incuor a normal intermediate in the DNA inbated at 0” and lysed (10.5% of the total jection process? To answer this question radioactivity) or untreated (10%) and in those incubated at 37” (11 and 12%, respec- the following experiment was done: 14Ctively). Then, the Pl pellets, resuspended labeled T5-E. coli complexes were prein the original volume, were treated with SYcrole conce”lrotion pancreatic DNase and recentrifuged; this 60% 40% c---) 20% test shows the same sensitivity to nuI 1 I I cleases for the DNA attached to the debris corn/fraction I I I at 0” (95%) as at 37” (91%), except perhaps 750 the FST fragment at 37”; the untreated bacteria have 53 and 78% of naked attached DNA at 0 and 37”, respectively. In 500 another experiment, the lysed bacteria were not centrifuged at low speed but imi 250 mediately analyzed according to the techi I I nique of Knippers and Stratling (1970). Figure 3 shows that the attached T5 DNA sedimented at the level of the so-called “DNA-cell envelope complex” extracted Fraction no. from bacteria infected at 0 and 37”; after FIG. 3. T5 DNA attachement analyzed as DNApancreatic DNase treatment of bacterial membrane complex. The technique of Knippers and debris before the ultracentrifugation, all Strltling (1971) was used to study the attachment of the radioactivity remained at the top of the T5 DNA to bacterial debris, without (0) and with gradient. Thus, it seems that there is ac- (0) DNase treatment.
IIi\!’\ 1;IIi : I-
STEP IN T5 DNA
pared at 0” in 4 ml of buffer and divided in two before centrifugation. The supernatants were removed and one pellet was covered with 2 ml of buffer and transferred to 37” on a slow rotary motor; the other pellet, covered with buffer, was kept at 0”. After 30 min of incubation, the two pellets, half-resuspended, were pipetted to achieve resuspension and then processed through the usual cycles of centrifugation-resuspension at 0”. Table 2 shows that in the sample incubated at 37” for the Pl pellet resuspension, the infected bacteria in P8 retained 8% of the radioactivity present in Pl; on the contrary, the sample kept at 0” retained only 1% of the Pl radioactivity in the P8 pellet. To verify that this 8% radioactivity corresponded well to the injected FST fragment, I extracted the DNA from infected bacteria in P8, and analyzed it on a neutral sucrose gradient: As Fig. 4 shows, I obtained a DNA segment measuring 8% in length of the total T5 DNA. Thus the exposed DNA attached at 0” may inject its FST part by a simple temperature transfer, as for the encapsided DNA. The attachment step actually appears to be an intermediate in the T5 infection process. DISCUSSION
The experiments described in this paper allow us to define a new, very early step of the T5 DNA injection process: In a situation where penetration of the FST-DNA segment inside the host cytoplasm has not yet occurred, the phage chromosome is attached to the infected bacteria as soon as it is released from the phage head through the tail. Such a situation was experimentally isolated in the injection process by simply incubating the bacteria at 0” in the presence of T5 phages in order to allow a normal adsorption while preventing FSTDNA injection (Fig. 1); then after a centrifugation which breaks away the upper part of some of the adsorbed capsids (Labedan et al., 1973), their DNA is found exposed and linked to pelleted bacteria. Thus it appears that there is, in the complex structure made of the adsorbed capsid and of the outer part of the cell envelope, a device able to attach the T5 DNA. The site of the DNA attachment
0” Pl P2 P3 P4 P5” P6 P7 P8
Radioactivity (%) at incubation temperature”
100 77 48 40 5.5 2 1 0.5
37 100 62 45 32 13.5 11 9.4 7.7
o The two pellets of bacteria, infected and centrifuged at 0”, were incubated either at 0” or at 37” before the subsequent cycles of centrifugation-resuspension. b After resuspension, the bacteria of the P4 pellet were treated with DNase before the next centrifugation leading to the P5 pellet.
may be characterized by the following properties: (1) It is free of host nucleases; (2) it is indifferent to the presence of calcium ions, thus differing from the site of fixation of the post-FST-DNA; (3) it is also insensitive to an EDTA treatment capable of releasing part of the lipopolysaccharide (Levy and Leive, 1968); (4) it is undamaged by the specific release of the cytoplasmic membrane from an extracted phage-envelope complex. A perplexing property of the T5 DNAattachment process is its relative weakness: When centrifuged T5-infected bacteria were gently resuspended during a long (at least 1 hr) time, the intact attached exposed DNA was spontaneously detached from bacteria. However, this attachment resists either repeated low-speed centrifugation (Fig. 1) or ultracentrifugation on sucrose gradient (Fig. 3). I have no explanation to resolve this paradox. The attached naked DNA may transfer its FST segment: This penetration was simply demonstrated by transfering to 37 the pelleted bacteria which were infected at 0”. Thus the attached DNA is normally fitted in the physiological injection process. The timing of this DNA attachment may be defined as prior to the steps of sensitization and densensitization to cal-
FIG. 4. Penetration of the naked attached FST’DNA. The P8 pellet (Table 2) was lysed according to Labedanetal. (1973). Of the extracted W-labeled P8 DNA (01,200 ~1 was deposited, with 2 pg of 3H-labeled T5 DNA (0) extracted by perchlorate, onto the same neutral sucrose gradient made and centrifuged as described in Labedan et 61.-(1973).
cium ions defined by Lanni (1961). Thus, in the external medium. Thus, under their conditions the phage receptor complex had this attachment appears as a transient state before the penetration of the FST- lost the DNA attachment property; it DNA segment and the subsequent tran- seems that a hypothetical cell component, which may be lost or inactivated during sient block of the chromosome injection. Recently, Scandella and Arber (1974) char- the T5 receptor extraction, could be necesacterized another earlier step of DNA sary for the attachment. Very recently, Saigo (1975) showed that phage injection: They described an E. coli there is in the T5 tail proximal end a mutant which adsorbs h phage normally but prevents its DNA injection by lack of a mechanism of DNA connection. In the inhost component triggering the DNA tact virion the DNA appears to be connected by its FST end; after an artificial release from the phage head. The aberrant T5 phages cannot attach ejection, the chromosome seems to be contheir chromosome to the infected bacteria. nected at the level of one of its singleIn effect, their DNA is ejected outside the strand interruptions. The DNA attachbacteria immediately after the phage ad- ment, described in this paper, was obsorption and secondarily degraded by host served even when the tail proximal end periplasmic endonuclease I (Labedan and was broken away by centrifugation and Legault-Demare, 1974). Thus, these phages appeared to be mediated by the FST end. are deficient in a component necessary to This DNA attachment to the cell surface allow DNA attachment in a site free of does not disagree with a possible connecnucleases. tion of the broken tail proximal end to Two hypotheses may be raised about the some other point of the DNA molecule. A DNA attachment mechanism: It is me- clear example of these two possible linkdiated either only by a constituent of the ages on the same DNA may be found in tail distal end or by the necessary interac- some electron micrographs published in a tion between some phage constituent and preceding paper (Plate III in Labedan et some structure of the cell envelope. The al., 1973): The naked uncoiled DNA was results described in this paper are not attached at its FST end to the bacteria and enough to make a clear choice between at the other end to the upper part of the these two hypotheses. However, Zaryb- broken capsid. Thus it appears that after nicky et al. (1973) showed that T5 adsorbed phage adsorption, the DNA injection proconto isolated receptors released their DNA ess implies a transport of the FST DNA
STEP IN T5 DNA
end from the connection point described by Saigo (1975) to the attachment point described in this paper. ACKNOWLEDGMENTS I wish to thank J. Legault-Demare for many helpful discussions during the course of this work. This work was supported by grants from the C.N.R.S. and the D.G.R.S.T. REFERENCES KNIPPERS, R., and STR~~TLING, W. (1970). The DNA replicating capacity of isolated E. coli cell wallmembrane complexes. Nature (London) 226, 713717. LABEDAN, B., CROCHET, M., LEGAULT-DEMARE, J., and STEVENS, B. J. (1973). Location of the first step transfer fragment and single-strand interruptions in T5stO bacteriophage DNA. J. Mo2ec. Biol. 75, 213-234. LABEDAN, B., and LEGAULT-DEMARE, J. (19741. Evidence for heterogeneity in populations of T5 bacteriophage. J. Viral. 13, 1093-1100. LANNI, Y. T. (1961). Invasion by bacteriophage T5.
III. Stages revealed by changes in susceptibility of early complexes to abortive infection. Virology 15, 127-135. LANNI, Y. T. (1968). First-step transfer deoxyribonucleic acid of bacteriophage T5. Bacterial. Rev. 32, 227-242. LEVY, S. B., and LEIVE, L. (1968). An equilibrium between two fractions of lipopolysaccharide in Escherichia coli. Proc. Nat. Acad. Sci. USA 62, 14351439. SAIGO, K. (1975). Tail-DNA connection and chromosome structure in bacteriophage T5. Virology 68, 154-165. SCANDELLA, D., and ARBER, W. (1974). An Escherichia coli mutant which inhibits the injection of phage A DNA. Virology 58, 504-513. SCHNAITMAN, C. A. (1971). Solubilization of the cytoplasmic membrane of Escherichia coli by Triton X-100. J. Bacterial. 108, 545-552. WEIDEL, W., KOCH, G., and BOBOSH, K. (1954). Uber die rezeptor-substanz fur den phagen T5. 2. Naturforsch. 9b, 573-579. ZARYBNICKY, V., ZARYBNICKA, A., and FRANK, H. (1973). Infection process of T5 phages. I. Ejection of T5 DNA on isolated receptors. Virology 54,318329.