Inactivation of urea-treated phage T4 by phosphatidylglycerol

VIROLOGY

41, 356-364 (1970)

Inactivation

of Urea-Treated LINDA

Department

Phage

T4 by Phosphatidylglycerol

BAUMANN,’ WENDY C. BENZ, A. WRIGHT, AND EDWARD B. GOLDBERG

of Molecular

Biology and Microbiology, Tufts University Medical School, Boston, Massachusetts 0211 Accepted February

20, 19YO

is inhibited Urea-treated phage T4 infection of Escherichia coli B,/4 spheroplasts by spheroplast debris. Almost all the inhibitory activity is due to phosphatidylglycerol (PG), one of t’he three major phospholipids of the cell. Neither phosphatidylethanolamine nor cardiolipin, the other major phospholipids, show significant inhibitory activity. PG from both Salmonella and Staphylococcus inhibits infection to the same degree. It has no effect on normal T4 infection of E. coli B cells. Electron micrographs of urea-treated phage preparations indicate that phosphatidylglycerol causes about 507; of the phage particles to lose their DNA and also brings about significant clumping of phage particles. These findings indicate that PG might either play a role as the urea-treated phage receptor on the spheroplast, surface or it might act as a trigger causing release of DNA from the phage. INTRODUCTION

Infection of Escherichia coli bacteria by phage T4 involves an initial interaction between the long tail fibers of the phage and the specific lipopolysaccharide receptor on the bacterial cell surface (Simon and Anderson, 1967; Wilson et al., 1970). III contrast, heat or urea-treated phage (UT4) which infect E. coli spheroplasts but not whole cells (Spizizen, 1957; Fraser et al., 1957; van Arkel et al., 1961; Wais and Goldberg, 1969; Baumann, 1969) require neither the long tail fibers nor the lipopolysaccharide receptors for infection. Wais and Goldberg (1969) have recently proposed that urea activates an attachment apparatus of the phage, distinct from the long tail fibers, which combines with a layer of the cell surface exposed by growth in penicillin or by lysozyme treatment. The phage genome is transferred via a DNase-insensitive bridge, and a normal eclipse and latent period ensue, leading to production of normal T4 phage. In an attempt to characterize the 1 Present address: Department of Microbiology, Universit,y of Hawaii, Honolulu, Hawaii. 356

spheroplast receptor, we have purified from spheroplast debris a substance which prevents infection of spheroplasts by UT4. We have characterized this substance as phosphatidylglycerol (PG). MATERIALS

AND

METHODS

Many of the Materials and Methods, including urea treatment of phage T4, have been described previously by Wais and Goldberg (1969); those not described previously are given below. T4Dac41, from Dr. A. Doermann, was the bacteriophage used in this work. E. coli B/4 and E. coli B40SuI/4 (from Dr. Strigini) spheroplasts were the host cells for UT+ infection. L-Broth contained 10 g of tryptone, 5 g of yeast extract (both Difco), 5 g of NaCl in 1 liter, adjusted to pH 7 with NaOH. Phosphate bu$er was 0.01 A1 sodium phosphate, pH 7.0. Tris HCl b@‘er was 0.01 M, pH 7.8. The following enzymes were used: pancreatic RNase and pancreatin from Sigma Chemical Co., pancreatic DNase and pronase from Calbiochem., salt-free lysoxyme from Worthington Biochemical Corp., penicillin-G from Sigma. PG from Staphylococczcs

aweus \v:ls a gift from Dr. William Lennarz. PG from SaZwo,lella ar~atunz was isolated as previously described (Dankert et al., 1966). Glucose-fructose-14C (200 IICi/prnole) was prepared by the method of Abraham and Hassid (1957). Penicillins spheroplasts: E. coli B,j4 cells grown in Pbroth were converted to spheroplasts by the method of Wais and Goldberg (1969). Lysozynze-EDT.4 sphereplasts were prepared from E. coZi B/-l- cells gro\\-n in P-broth to OD,,, = 0.5 (1.4 X lo* cells/ml). The cells were harvested by tenbrifugation, \\-ashed in Tris buffer, and resuspended in one-third volume of Tris buffer containing 13 5%sucrose. One milliliter of cells was incubated for 5 min at 37” with 0.01 ml of a mixture containing 0.04 pmole of EDTA, 0.05 mg lysozyme (456 units) and 1.4 kmoles of Tris buffer. The resulting spheroplasO suspension was mixed with 1 ~1 of SO m-l1 MgSO, and kept at 0” until used. E. coli B40SuI/d lysozyme spheroplasts were prepared by concentrating cells G-fold with Tris buffer containing 1.55% sucrose, and 0.02 ml of lysozyme-EDTA mixture w-as added per milliliter of cells. The cells were mixed w-i& 4 ~1 of 0.2 N l\IgSOI. Lysetl spheroplast debris (LYSD). E. coli B/4 cells, grown and harvested as described above, were \\-ashed twice with Tris buffer then resuspended in 1,000 volume of Tris buffer. The concentrated cell suspension n-as incubat#ed with 0.37 pmole of EDTA and 2.25 mg (20,473 units) of lysozyme per milliliter for 5 min at 37”. Spheroplasts lyse as they form since no sucrose is present. After addit8ion of 1.9 ~1 of 0.2 31 SlgSOa per milliliter of cell suspension, the mixture was sonicated for 3 min at 0” giving 100 X LYSD. Fivefold (.5X), IO-fold (10X) and 20-fold (20X) concentrated LYSD were prepared by diluting 100 X LYSD with Tris buffer. “C-labeletl LYSD: E. coli B/4 cells were labeled by adding 0.3 mCi of glucose-fructose-14C mixture to 500 ml of L-broth culture at ODeO,,= 0.5. The culture was chilled and 100 X LYSD was prepared in the usual manner. Butanol extractiotb of phospholipids from LYSD. Equal volumes of LYSD, distilled water, and n-butanol were vigorously mixed, and the butanol layer was removed after centrifugation. The aqueous phase was re-

extracted once or twice (as indicated) with an approximately equal volume of wbutanol. The combined butanol extracts were washed 3 times \vith about one-third the volume of water for each mash. In the experiments described in Tables 1 and 2 butanol was removed by dialysis but with poor recovery (lo-20 %). In subsequent experiments solvent was removed by evaporation as described below. DEAE-cellulose chromatography of phospholipids was carried out as described by Dankert et al. (1966). The washed butanol extract of ‘*C-labeled 100 X LYSD (18 ml) was applied to the DUE-cellulose column. Three major radioactive peaks were obtained corresponding to phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and cardiolipin (CL). The column fractions corresponding to these peaks were combined and concentrated under vacuum to remove solvent. Ammonium acetate, which was present in the PG and CL fractions, was removed by dissolving the residues in 1.5 ml of n-butanol and washing 3 times with OS-ml portions of distilled water. Each of the combined peak fractions ~vas concentrated to dryness and redissolved in small volumes of solvent, 0.1 ml methanol in the case of PC, and CL and 0.3 ml of CHQ-methanol (2: 1) in the case of PE. Each fraction was mixed with 2 ml of Tris buffer, and solvent was removed b! concentrating to 1.5 ml under a strea,m of nit’rogen at 40”. The recoveries of the 3 fractions calculated from radioactivity were PI? 5ci%, PG SS%, and CL 100%. Control samples without phospholipids did not inhibit’ phage production in the UT+spheroplast system. The aqueous suspensions of phospholipids in Tris buffer were used directly to t’est their inhibitory potential. The total phosphate present in each sample was determined (Ames, 1966) and used as a measure of phospholipid concentration. Puri’catio~l of PC; by paper chromatography n-as carried out on glass fiber paper impregnated with silica gel (Mallinckrodt Chrom AR500) using chloroform-methanol-water (75 : 25 : 2.7) as solvent. 14C-Labeled PG (0.13 pmole) was chromatographed in this system, and radioactive spots were localized by radioautography. The appropriate regions of the chromatogram were cut out,

35s

ET AL.

BAUMANS

eluted with methanol (0.1-1.2 ml), diluted with 1.5 ml of 0.01 M Tris buffer, then concentrated to 1.0 ml under a stream of nitrogen at 40”. These samples were used directly for measuring inhibitory potential. Assay for inhibitory potential. Four-tenths milliliter of P-broth, 0.2 ml of inhibitor in Tris buffer, and 0.1 ml UT+ added in this order, were preincubated for 5 min in a SO-ml Erlenmeyer flask at 37”. After preincubation, 0.5 ml of B/4 spheroplasts was added and the mixture was incubated at 37” for 80 min. The incubation was terminated by addition of chloroform, and the phage yield was determined by titration. In early experiments (Fig. l), there was no preincubation period. In the spheroplast infective center assay for inhibitory potential, CT+ was preincubated for 10 min at 37” with PG

in 0.01 IV phosphate buffer, pH 7. Then 0.01 to 0.1 ml of the appropriate dilut,ions of UT+ were added to a mixture of 0.5 ml of fi. coli B40SuI/4 lysozyme-EDT-1 spheroplasts diluted 1: 1 w-lth P-broth. Spheroplast infective centers w-err plated after 10 min at 37” by the method of Wais and Goldberg (1969). Elect~orz mboscopy. A UT4 preparation containing 3 X 10” particles per milliliter was applied to a Formvar grid and negatively stained with 2 ‘% silicotungstic acid adjusted to pH 6.9 with SaOH. PG treatment was carried out by incubating equal volumes of UT@ (4 X 10”’ particles per milliliter) and PG (70 mpmoles/ml) in 0.01 31 Tris buffer for 5 min at 37”. The mixture was applied to a Formvar grid and stained as described above.

20

PHAGE YIELD X10+

W4 PENICILLIN (CELL

SPHEROPLAST DEBRIS x lo-7 EOUlVALENTS/ML.)

FIG. 1. Relation between phage production and penicillin spheroplast debris concentration. Penicillin spheroplasts were concentrated by centrifugation (39,000 g for 10 min), resuspended in one-half volume of Tris buffer, and incubated with 200 pg/ml DNase and RNase for 20 min at 37”. The debris was washed twice with 0.9% NaCl and once with Tris buffer and resuspended at lOlo cell equivalents per milliliter in Tris buffer. A mixture of 0.5 ml of penicillin spheroplasts, 0.4 ml of Tris buffer, 0.2 ml of various concentrations of the penicillin spheroplast debris, and 0.1 ml UT+ was incubated at 37” for 80 min, CHCla was then added, and the phage yield was determined by titration.

ISACTIVATION RESULTS

AND

OF T.RES-Tl:I.;hTI?l>

PHAGE

INHIBITORY

POTENTIIL

DISCUSSION

Infection of 13. coli H/4 spheroplasts by UT@ is inhibited by spheroplast debris. When penicillin spheroplasts mere lysed and washed in Tris buffer and the washed debris TVRSadded t’o a UT+-spheroplast mixture, the production of phage was inhibited as shown in Fig. 1. The inhibition of phage production depends on t’he amount of spheroplast debris added although the relationship is not linear. In subsequent experiments we found that) debris from both penicillin and lysozymr spheroplnsts inhibits infection of E. cob’ B/-l spheroplnst’s by UT+, however, the latter was easier to prepare and the results obtained were more reproducible. Therefore, we used lysozyme sphrroplast debris for all subsequent experiments. To determine whether inhibition by L‘1XD was due to interaction with UT4, LYSD, and UT4 were preincubated in the absence of spheroplasts. Figure 2 shows that the inhibit,ion of phage production is great13 enhanced by preincubation of UT4 with L‘1XD. These results indicate that LYSD acts directly on UT+ to inhibit phage production. In order to characterize the inhibitor in the sphrroplast debris we tested the effect, of various physical, chemical, and enzymatic treatments on inhibitory potential. Neither

TIME IN MINUTES FIG. 2. The effect of time of preincubation of UT+ and LYSD on phage yield. Mixtures of 0.1 ml of UT4, 0.2 ml of, LYSD and 0.4 ml of P-broth were preincubated for the times indicated before adding 0.5 ml of spheroplasts to t,he mixture. Phage production was measured as described in the legend t)o Fig. 1.

TABLE

Experiment

Control 1

2

3

1

OF LYSD

~ Additions

?r’oue LY81) LYSD LYSI) LYSD LYSI) LYSD LYSI) LYYI) ~ LYPI) / LYSI)

359

T4

AFTER

Relative yield of phage

Treatment

SOME Sonicatinn Heat Freeze-thaw None Pronase Trypsin None Butauol

1 Petroleum

PHYSI-

ether

-1.0 0. 1 0.002 0 .2 0.1 0.04 O.OF 0.07 0.0-l 0.7 0.4

n After treatment, the LYSD samples were tested for inhibitory activity in the UT+spheroplnst assay system. Sonicationt 5 X LYSD (0.52.0 ml) sonicated for 15 set at 0”. Heat: 5 X LYSD heated in a sealed tube in a boiling-water bath for 10 min. Freeze-thaw: 10 X LYSD, frozen (to -20’) and thawed (at 37”) three times. I’ronase: 20 X LYSD was treated with 5 @g/ml of promise for 30 min at 37”. Pronase was inactivated before assay by heat treatment. l’rypsin: 5 X LYSD was treated with 100 pg/ml of trypsin for 15 min at 37”. Trypsin was inactivated before assay by adding 200 ~g/ml of trypsin inhibitor. Rutanol e.draction: 0.G ml of 20 X LYSD was extracted twice with butalLo1 as described in Materials and Methods. The aqueous layer was dialyzed overnight, against Tris buffer to remove solvent. After dialysis, samples were sonicated for 15 set t,o disperse any material that might be aggregat,ed. Petroleutn ether: 0.0 ml of 20 X LYSD was extracted with 0.G ml of petroleum ether by the same procedrlre as used for but,anol extraction. Controls, without added LYSD (Yield = l.O), were carried out in each experiment. The control phage yields after incttbation varied from 3 X lo6 to 2.5 X 10’ PFU,‘ml. Before incubation, the background of normal phage in t.he urea-treated preparation was less than 100 PFU/ml.

RSase nor DBase at’ concentrations of about 1 mg/ml had any effect on inhibitory potential. Table 1 shows that the inhibitory potential \vas unaffected by treatment with heat, freeze-thawing or proteolytic enzymes. Sonication greatly enhanced inhibitory potential whereas solvent extraction decreased it. These properties implicated a lipidlike material as the inhibitor. This was con-

360

BAUMANN TABLE

INHIBITORY

POTENTI.\L

TABLE

2

0F R~T~NoL-E;~TR~C,~ED

LYSDU Water phase From

From

LYSD + + Tris + -

Butanol phase

-

buffer

+ + (control) +

El’ AI,

INHIBITORY LIPIDS

Phospholipid (rnp moles)

Relative yield of phage

0.75 0.02 0.01 1.0 1.0

a The water phase was prepared from 0 6 ml aliquots of 100 X LYSD extracted as described in Table 1. The butanol phase was prepared by washing the combined bntanol layers three times with 0.5 ml of water and dialyzing overnight against Tris buffer to remove the solvent. Thf. nondialyzable material from each phase was assayed for inhibitory potential. For control ex tractions, LYSD was substituted by Tris buffer. The yield of phage when no additions were made (LYSD, water phase, or butanol phase) was lo7 PFU/ml (~1.0).

firmed by testing butanol-soluble material from LYSD for inhibitory activity, after removing the solvent. Table 2 shows that this material significantly reduces the phage titer, whereas the material remaining in the water phase retains very little inhibitory activity. The combined material from the butanol and water phases did not show a significantly greater inhibition than the material from the butanol phase alone which suggests that most of the inhibitory activity is present in the butanol phase. A butanol extract of 14C-labeled LYSD was further purified by chromatography on a DEAE-cellulose column (Dankert et al., 1966). Three major 14C-labeled fractions were obtained, corresponding to phosphatidylethanolamine (PE), phosphatidylglycerol (PG), and cardiolipin (CL). In bacteria, phospholipids are found almost exclusively in the cell membrane (Rothfield and Finkelstein, 1968). Table 3 shows that almost all the inhibitory activity was found in the PG fraction. At the higher concentration the material in the CL fraction gave significant inhibition but was, on a molar basis, only 0.5% as active as the material in the PG fraction. Inhibition could be due to CL it-

Control Phosphatidylethanolamine Phosphatidylglycerol Cardiolipin

3

POTENTIAL OF PURIFIM) coli FROM Escherichia

Phage yield (PFU/ml)

0 6.8 13.5

5.3 x 7.8 x 0.8 x

6.3 12.5 7.9 15.8

1.4 6.0 2.7 1.1

106 10” 10”

x IGj x 10” x 106 x 106

PHOSPHO-

I!,‘* Relative yield 1.0 0.94 0.82 0.017 0.0007 0.33 0.13

n Aqueous suspensions of phospholipids from E. coli B/4 were tested directly for their inhibitory pot,ential in a UT@spheroplast mixture as described in Materials and Methods.

self, to a minor component in the CL fraction, or to a breakdown product of CL. No inhibitory activity was present in the PE fraction, which is the major lipid fraction from these cells. Chromatography of the 14C-labeled PG fraction on glass fiber paper impregnated with silicic acid gave two radioactive components. The region corresponding to PG contained 90% of the recovered 14C and all the inhibitory activity. Ten percent of the radioactivity was found at the solvent front (probably corresponding to a fatty acid breakdown product of PC), and it had no inhibitory activity. No radioactivity nor inhibitory activity was found elsewhere on the chromatogram. Thus PG, the major component of spheroplasts with inhibitory activity, is highly specific in the inactivation of UT+ The identity of PG as the active component is supported by the finding that PG samples from Salmonella anafum and from Staphylococcus uuwus also inactivate UT4, and furthermore do so to the same degree as those from E. coli. Log phase cells of E. coli and Salmonella contain a significant proportion of monounsaturated fatty acids in their PG (Kanemasa et al., 1967), whereas S. aureu.s PG contains mostly branched and straight-chain saturated fatty acids (Macfarlane, 1962). Thus, the specific fatty acid component of the PG is not critical for inhibitory activity. PE and CL have the same fatty acid composition (Kanemasa et al., 1967) in E. coli B as PG but either do

FIG. 3. (a) Urc:t-treated T-l phage showing frdl heads atld contracted sheaths. (b) Urea-treated T4 phage following treatment with Phosphatidylgl?-cerol. Approximately 50c; of the phage have lost theit IINA, and there is a significant cltm~ping. Negatively stained with silirot ungstic acid. The bars ill thesca micrographs reprcserrt 1000 ~1. Xl

362

BAUMANN

not inhibit or inhibit to a much lower degree. The specificity of PG must therefore depend upon parts of the molecule other than its fatty acid components. It is possible that the level of inhibition by phospholipids is affected by specific micelle formation in aqueous solvents. It is also possible that phospholipids other than PG may be active in vivo.

Electron micrographs of UT+ and PGtreated UT+ are shown in Fig. 3. The main feature of the UT+ particles (Fig. 3a) is that they all have contracted sheaths but still retain their DNA. PG treatment has t\vo principal effects: it causes many of the particles to lose their DNA and also causes the phage to form clusters (Fig. 3b). Similar

ET AL.

clusters occur in the case of untreated UT4 samples, but they are much less frequent. Formation of clusters may reflect attachment of UT+ to isolated membrane receptors and phospholipid micelles. PG catalyzes loss of DNA from approximately 50% of the UT4 particles. We measured UT+ inactivation and DKA release at various PG concentrations to see whether or not these effects are related (Fig. 4). At low PG concentrations there is significant DKA release without UT4 inactivation. In contrast, at higher PG concentrations, marked inactivation of UT4 is accompanied by only small increases in DKA release. Approximately SO-60 7%of the DNA is released at the higher PG concentra-

100

60

70

70

60

60

50

50

X INHIBITION OF INFECTIVE CENTERS

X DNA RELEASED 40

40

30

30

20

20

ID

IO

0

0 0

9

18

27

36

45

PG CONCENTRATION : ( mu MOLES/ML. UT$ I

FIG. 4. R.elation between UT+ inactivat,ion and DNA release caused by PG from Salmonella anatunl. 3WLabeled UT+ was preincubated with the indicat,ed concentrations of PG for 10 min at 37”. Inhibition of infective center formation was measured using the spheroplast infect.ive cent.er assay (Materials and Methods). DNA release was measured as perchloric acid soluble radioactivit,y produced by t,reating samples with DNase (O.OGmg/ml) for 10 min at 37”. After acid precipitation wit,h cold 37, perchloric acid, radioactivity in the supernat,ant was measured in a Nuclear Chicago gas flow low background counter. 0-0, Infective centers; O-----O, I)i%A release.

INACTIVATION

60

OF UREA-TREATED

PHAGE

-

T4 -

60

-

50

-

40 % DNA RELEASED

UT0

, 2

1 15

IO

5 MINUTES

30

-

20

-

IO

CONTROL

o-0

-

0

OF INCUBATION

Flo. 5. Time course of UT+ inactivation and DNA release caused by PG. Inhibition was tested in the spheroplast infective center assay. DNA release was measured as described in the Fig. 4 legend. The concentration of PG used was 0.018 pmole/ml, which gave about 50V,G inactivation in Fig. 4. @---a, Infective centers; O-----O, DNA release. tions,

confirming

the

electron

microscopic

observations. The rate of inactivation of UT4 by a PG concentration which gave 50 % inactivation in the previous experiment is shown in Fig. 5. The reaction is complete in less than 5 min. DNA release is even more rapid under these conditions. DNA release involves a significant fraction of the UT4 particles, whereas viable UT+ occur at a frequency of only one particle per thousand. Thus, it appears that DNA is released from most UT4 more easily than from the small fraction of infective UTC$. Inactivation of infective UT+ at elevated PG concentrations may be caused by DNA release or cluster formation. Thus it is possible that only those phage which do not release their DXA at low I’G concentrations can survive the initial contact with the spheroplast surface and form a fruitful junction. On the other hand, infective UT4 may also possess a I’G sensitive factor required for infection, in addition to that stabilizing them to release of DNA. Positive evidence for such a factor would require measuring attachment of infective UT4 to spheroplasts, a task at which we have not yet succeeded We suspect. that UT+spheroplast interaction

is related

to one of the stages

in nor-

mal T4 infection subsequent to the initial long-tail-fiber interaction with lipopolysaccharide. A different receptor must be present on the spheroplast surface for interaction with UT+ Although inhibition of UT4 infection of spheroplasts by PG is highly specific, we do not yet know whether it is gratuitous or whether 1’G is in fact a component of t’liis receptor. ACKNOWLEDGMENT We wish to thauk Dr. W. 1). Belt for his help with the electron microscopy and for the use of the electron microscope. This research was supported by Grants GM13511, GM-15837 from the National InstitlLtes of Health, GB 5923 from the National Science Foundation, and 68-833 from The American Heart Association. One of IIS (L. B.) was a predoctoral fellow of the U.S. Public Health Service, and E.G. is a career development, awardee of the U.S. Pllblic Health Service GM-7567. REFERENCES and KISSID, W. %. (1957). Synthesis of labeled carbohydrates. 1n “Methods in Enzymology” (S. P. Colowick and N. C. Kaplan, eds.) TVol. 4, pp. 489-494. Academic Press, Xew York. AMES, B. N. (19fiO). Assay of inorganic phosphate,

ABR.IH.\M,

P.,

364

BAUMANN

total phosphate and phosphatases. In “Methods in Enzymology” (E. F. Neufeld and 17. Ginsburg, eds.), Vol. 8, pp. 115-118. Academic Press, New York. BAUMANN, L. (1969). Study of t.he mechanism of urea-treated T4 phage infection of E. coli spheroplasts. Master’s Thesis, Tufts University. DANKERT, M., WRIGHT, A., KELLEY, W. S., and ROBBINS, P. W. (1966). Isolation, purification and properties of the lipid-linked intermediates of O-antigen biosynthesis. Arch. Rio&em. Biophys. 116, 425435. FRASER, I>., MAHLER, H. R., SHUG, A. L., and THOMAS, C. A., JR. (1957). The infection of subcellular Escherichia coli, strain B, with a DNA preparation from T2 bacteriophage. Proc. Natl. Acad. Sci. U.S. 43, 939-947. Kh~~~as.4, Y., AIUM.\TSU, Y., and NOJIM.~, S. (1967). Composition and turnover of the phospholipids in Escherichia coli. Biochim. Biophys. Acta 144, 382-390. MACFARL.INE, M. G. (1962). Lipid components of Staphylococcus aureus and Salmonella typhimurium. Biochem. J. 82, 4OP-41P.

El’

AL.

ROTHFIELD, L., and FINKIILSTEIN, A. (1968). Membrane biochemistry. Ann. Rev. Biochem. 37, 463-496. SIMON, L. D., and ANDERSON, T. F. (1967). The Infection of Escherichia coli by T2 and T4 Bacteriophages as seen in the elect,ron microscope. I. Bttachment and penetration. II. St,ructure and function of the base plat,e. Virology 32, 279-297, 298-305. SPIZIZEN, J. (1957). Infection of protoplasts by disrupted T2 virus. Proc. Natl. Acad. Sri. U.S. 43, 694-701. v.4~ ARKEL, G. A., 17.4~DE POL, J. H., and COHEN, J. A. (1961). Genetic recombination and marker rescue of urea-disrupted bacteriophage T4 in spheroplasts of E. co&. Virology 13,546-548. WAIS, A., and GOLDBERG, E. B. (1969). Growth and Transformation of T4 Phage in E. coli B/4, Salmonella, Proteus, and Serratia. Virology 39, 1533161. WILSON, J. H., LUFTIG, R. B., and WOOD, W. B. (1970). The interaction of bacteriophage T4 tail fiber components with purified E. coli lipopolysaccharide. J. Mol. Biol., in press.