Formation of phage T1 concatemers by the RecE recombination pathway of Escherichia coli

VIROLOGY

135,200-206

(1934)

Formation of Phage Tl Concatemers by the l+cE Recombination Pathway of Escherichia co/i J. C. PUGH’

AND

D. A. RITCHIE’

Department of Genetics, University of Liverpool,

Liverpool

L69 SBX, England

Received October 20, 1983;accepted March 2, 1981 Infections of nonpermissive (sup’) Escherichia coli by Tl phage with amber mutations in either gene 3.5 or gene 4 exhibit a variety of defective phenotypes, including premature arrest of Tl DNA synthesis, failure to make concatemeric DNA, formation of an abnormal DNA replication intermediate, failure to package phage DNA, and reduced genetic recombination. The lethal effect of gene 3.5 or 4 mutations is suppressed when the sup’ bacteria express the RecE recombination pathway. This RecE suppression occurs by partial restoration of the capacity to make concatemeric molecules and partial reversal of the DNA arrest defect which, in turn, leads to the formation of viable progeny. Infection by Tl+ or by mutants defective in any of the four DNA synthesis genes (genes 1, 2, 3.5, and 4) inhibited the ATP-dependent exonuclease present in uninfected cells (presumably the RecBC enzyme, exonuclease V). Extracts from Tl+ infections also showed increased levels of an ATP-independent exonuclease activity which was absent from gene 4 mutant extracts. It is concluded that gene 4, together with gene 3.5, specifies an activity related to that of the RecE exonuclease VIII and essential for Tl concatemer formation and recombination.

preinfected with phage X. X did not complement amber mutations in any other Tl During infection of nonpermissive (sup’) genes known at that time, and complecells by amber mutants in Tl genes 3.5 and mentation of am23 did not occur with X 4, phage DNA sunthesis is arrested pre- mutants defective in the reda and redp maturely, formation of concatemeric DNA general recombination functions. Subsemolecules is inhibited, and genetic recom- quently, we showed that am23 phage could also replicate in E. wli sup’ cells expressing bination is drastically reduced (Figurski and Christensen, 1974; Ritchie et ak, 1980; the RecE recombination pathway as did Ritchie and Joicey, 1980; Pugh, 1982). The am201, a mutant in the newly identified DNA synthesised during the early stages Tl gene 3.5 (Ritchie et aL, 1980). The product of the RecE pathway is of gene 3.5- and gene 4- infections consists of molecules which sediment more slowly thought to be exonuclease VIII, an ATPthan Tl virion DNA, as a result of having independent, 5’-specific, double-strandedspecific DNA exonuclease with no activity lost approximately 12% of their terminal sequences (Pugh and Ritchie, 1984). These at nicks or gaps (Gillen et ak, 1981; Sadowski, 1982). Exonuclease VIII has chemshorter molecules are stable throughout infection. Christensen (1976) made the in- ical properties similar to that of the X exoteresting observation that Escherichia coli nuclease, the product of the X reda gene, sup’ cells will support the replication of and is made by bacteria harbouring the the Tl gene 4 amber mutant, am23, when cryptic lambdoid prophage (Rat) (Radding et aL, 1967; Kaiser and Murray, 1979). The RecE pathway is controlled by the sbcA ’ Present address: Department of Molecular Biology, mutation, which restores the capacity for University of Edinburgh, Edinburgh, EH9 3JR, Scotgenetic recombination to E. co&i recB recC land. ‘To whom reprint requests should be addressed. mutants, presumably by allowing the conINTRODUCTION

9042~6322/&l $3.00 Copyright All rights

0 1984 by Academic Press, Inc. of reproduction in any form reserved.

200

RecE FORMATION

OF PHAGE

Tl

CONCATEMERS

201

stitutive expression of exonuclease VIII 37” in 100 ml supplemented M9 medium (Barbour et &, 1970; Kushner et a& 1974). (Ritchie and Joicey, 1978) to 1 X lo8 cells/ Thus, the loss of E.coli recombination ml. [3H]TdR was added (2 &i/ml), and infunctions by mutation of the recB or recC cubation continued to a density of 3 X lo* genes is compensated for by the expression cells/ml, when the cells were harvested by of the RecE pathway which occurs when centrifugation and suspended in 2 ml 27% sbcA mutations arise. Wild-type bacteria w/v sucrose in 1X SSC (0.15 MNaCl, 0.015 do not express the RecE pathway. The ca- Msodium citrate). The cells were incubated pacity for the RecE and X recombination for 5 min at 4’ with lysozyme (0.5 mg/ml), pathways to substitute for the Tl gene 3.5 when the volume was made up to 38 ml and 4 recombination functions suggests with 27% sucrose containing protease K (2 that these Tl genes also code for an exo- pg/ml). After incubation at 37” for 15 min, nuclease with similar biological properties. the cells were lysed with SDS (1% final In this paper we investigate the ability concentration) by incubation overnight at of the E. coli RecE pathway to restore nor- 30”. Protein was removed by phenol exmal DNA replication in sup0 cells infected traction (Thomas and Abelson, 1966) and with Tl gene 3.4 and 4 amber mutants, and the aqueous phase, adjusted to 1 M in NaCl, show that Tl gene 4-infected cells are de- was chilled in ice for 15 min and centrifective for a Tl-coded exonuclease. fuged at 12,000 g for 10 min to sediment the SDS pellet. The supernatant, after exMATERIALS AND METHODS tensive dialysis against 1X SSC, was Phage and bacteria. The origins and treated with pancreatic RNase (60 pg/ml) properties of the phage Tl wild type and for 1 hr at 37”, again phenol extracted, and amber mutants, am23 (gene 4) and am201 dialysed exhaustively against 0.1X SSC. (gene 3.5), have been described previously This DNA was used as a substrate for nu(Ritchie and Joicey, 1980, 1983; Ritchie et clease assays within 2 weeks of preparauL, 1980; Michalke, 1967; Figurski and tion. Radioactive labelling of intracellular Tl Christensen, 1974). E. coli strains KB-3 (recB+ s&A+ supE) and B (recB+ sbcA+ DNA. Replicating Tl DNA was labelled sup’) were the standard permissive and with [3H]TdR using pulse and pulse-chase nonpermissive hosts, respectively, for the conditions as described by Ritchie and JoiTl amber mutants. Strain B41 (endoII cey (1978, 1980) and Pugh and’ Ritchie sup”), a gift of the late Dr. Mary Lunt, was (1984). Infected cells were lysed either with used to prepare cell extracts for assaying SDS, which disrupts both cells and phage nuclease activity. Strain JC5412 (recB21 particles, or with sodium deoxycholate sbcA8 sup’), isolated by Barbour et aL (DOC), which disrupts cells but not phage (1970), was kindly provided by Dr. K. K. particles. Preparation of cell-free extracts. Nuclease Kaiser. Strain B (thy- sup”) was isolated in this laboratory by the trimethoprim se- activities were determined with cell-free lection method (Miller, 1972). extracts by the method of Barbour and Preparation of 14C-labelled Tl DNA. A Clark (1970). Cultures (16 ml) were chilled culture of KB-3 was grown in TCG medium by pouring over frozen TCG medium, (MacHattie et a& 1967), with aeration, at washed in saline buffer (15 mMNaCl), and 37” to 1 X 10’ cells/ml. [14C]TdRwas added resuspended in 0.4 ml ice-cold 25% w/v su(0.2 &i/ml), and growth continued to a crose in 10 mM Tris, pH 8. EDTA, pH 8, titre of 3 X 10’ cells/ml, when the cells was added to a concentration of 10 mM and were infected with Tl+ phage (5 phage/ lysozyme at 0.1 mg/ml, and the mixture cell) and the culture was aerated until lysis. was left on ice for 2 min, followed by the Phages were purified, and the DNA was addition of MgS04 and Brij-58 to give final extracted with phenol as described by concentrations of 50 mM and 0.5% w/v, Ramsay and Ritchie (1980). respectively. The suspension was mixed Preparatim of sH-labelled E. coli DNA. well, left on ice for 60 min, and then cenE. coli strain B (thy- sup’) was grown at trifuged at 10,000 Q for 15 min at 4O. The

202

PUGH AND RITCHIE

supernatant was removed and used immediately. Nuclease assays. To 0.2-ml cell extracts was added 2-5 pg %I-labelled E. co& DNA (l@-105 cpm/pg) and 0.1 ml 2.5 n& ATP, where appropriate, and the volume was made to 0.5 ml with 10 mMTris, pH 8. The reaction mixture was incubated at 37”, and O.l-ml samples were removed at intervals into 0.5 ml ice-cold TCA (5% w/v) containing 0.5 mg bovine serum albumin as carrier. After 10 min on ice the mixture was centrifuged at 4’, and the total supernatant was added to 4 ml scintillation cocktail (5.5 g PPO, 0.1 g POPOP, 333 ml Triton X-100,667 ml toluene) and counted for radioactivity. Sucrose gradknt centrt&gation. These methods have been previously described (Ritchie and Joicey, 19’78;Pugh, 1982; Pugh and Ritchie, 1984). RESULTS

Replication of am23 and am201 DNA sup0 RecE’ Cells

in

A comparison of the overall pattern of DNA synthesis in sup’ and sup’ RecE+ cells showed that, in the sup’ RecE+ host, the arrest of DNA synthesis was less abrupt than in the sup0 host (data not shown). This pattern of DNA synthesis, which was intermediate between that of sup’ and sup+ bacteria, suggested that a good proportion of am23 and am201 DNA molecules replicate normally in sup’ bacteria with a functional RecE pathway. The host RecE pathway therefore restores, albeit partially, the late phase of Tl DNA synthesis which is arrested in sup’ cells. A short pulse of [3H]TdR was given to cultures of supE, sup’, and sup’ RecE+ cells from 9 to 10 min after infection with Tl am23, and samples were removed and lysed with SDS immediately after a 2-min pulse and also after a further lo-min chase period. Intracellular DNA was analysed by sedimentation through neutral sucrose gradients (Fig. 1). During permissive infection the sedimentation profile mimics that of wild-type phage with the rapid formation of fast-sedimenting concatemeric DNA molecules (Fig. lA), which were pro-

16-O

20

80

"b 800

160 0 mb

;

x

E I2 &O

B u

eo

,'

2 0

0

40

0

8.0

10

20

0

10

20

0

FRACTION NUIBER

FIG. 1. Sedimentation analysis of Tl am23 intracellular DNA labelled in a pulse-chase experiment. Bacterial cultures were infected at 30’ with phage, and the intracellular DNA was labelled by the addition of [8HpdR (10 &i/ml) to the growth medium at 8 min after infection. At 10 min after infection, excess unlabelled thymidine was added, and incubation continued until 20 min. Samples were removed at 10 (pulse) and 20 min (pulse-chase) after infection, lysed with SDS, and sedimented through neutral sucrose gradients at 40,000 rpm for 85 min at 10” (A) KB-3 (recB+ s&A+ ezqE) cells, lo-min sample; (B) JC5412 (recB21 &A8 8up”) cells, lo-min sample; (C) B (recB+ &A+ sup”) cells, lo-min sample; (D) KB-3 cells, 2Omin sample; (E) JC5412 cells, 20-min sample; (F) B cells, 20-min sample. Arrows show the position of “Clabelled Tl+ virion marker DNA sedimented in the same gradient. Sedimentation is from right to left.

cessedduring a chase into monomer-length molecules cosedimenting with Tl+ virion marker DNA (Fig. 1D). No concatemers were formed during infection of sup’ cells, and the replicating DNA was synthesised as stable molecules sedimenting slightly more slowly than virion DNA (Fig. lC, F), as reported by Pugh and Ritchie (1984). The DNA molecules synthesised in sup’ RecE+ cells show an intermediate sedimentation pattern (Fig. lB, E), with fastsedimenting concatemers being formed, but in a smaller proportion and with a reduced average size compared with sup+ cells (Fig. 1B). At the same time, there is clear evidence of the shorter DNA molecules characteristic of sup’ infections. Furthermore, following the chase period,

RecE FORMATION

OF PHAGE

the peak of radioactivity occupies a position intermediate between that observed for the sup+ and sup” infections, pointing to a mixed population of molecules containing both virion-length DNA and the shorter molecules typical and nonpermissive infections (Fig. 1E). A parallel experiment with Tl am201 phage gave very similar results. These results suggest that the concatemers formed during am23 and am201 infection of sup’ RecE+ cells are processed normally, while those monomers which fail to recombine into concatemers are shortened by the same mechanism that is responsible for degrading the ends of replicated molecules during nonpermissive infection (Pugh and Ritchie, 1934). Packaging of am.23 and am201 DNA from sup’ RecE+ E. c&i Figure 2 shows the result of an experiment designed to measure the efficiency with which am23 intracellular DNA was packaged into phage particles. Intracellular DNA was pulse-labelled with [3HJTdR from 10 to 12 min after infection, and the label was chased with nonradioactive thymidine until 19 min, when the cells were lysed with sodium deoxycholate, which does not disrupt mature phage particles. Samples were sedimented through sucrose gradients to distinguish packaged from unpackaged radioactive DNA. Under permissive (sup’) conditions of infection about 30% of the incorporated label was present as phage particles (Fig. 2A)-this figure is very similar to that observed with wildtype phage infections (Ritchie and Joicey, 1978). By contrast, only about 1% of the label from nonpermissive (sup”) infection was observed at the phage peak (Fig. 2C), a result which reflects the requirement of a concatemeric DNA substrate for efficient packaging. The lysates from sup” RecE+infected cells showed an intermediate value, with approximately 7% of the labelled DNA sedimenting as virus particles (Fig. 2B). Parallel experiments with am201 mutant phage gave essentially the same result. These results once again demonstrate that the substitution of the Tl gene 3.5 and

Tl

203

CONCATEMERS

10 FRACTION

20 NUMBER

FIG. 2. Sedimentation analysis of DOC lysates of Tl orr&%infected cells from pulse-chase experiments. Bacterial cultures were infected at 30” with phage, and the intracellular DNA was labelled by adding 5 &i/ml [‘HJTdR to the medium at 10 min. At 12 min after infection, excess unlabelled thymidine was added and incubation continued until 19 min, when the cells were lysed with DOC. Samples were sedimented through neutral sucrose gradients at 21,000 rpm for 45 min at 10”. E. coli host strains were (A) KB-3; (B) JC5412; and (C) B. Arrows show the position of i*Clabelled Tl+ phage particles sedimented in the same gradient. Sedimentation is from right to left.

gene 4 functions by the E. coli RecE function is only partial, and show that two factors contribute to this incomplete substitution. (i) Fewer concatemers are produced, and those that are made are shorter than in either wild-type infection or in am23 or am201 infection of permissive cells (Fig. 1). (ii) A large proportion of the DNA is replicated as the shortened form of molecule, characteristic of nonpermissive infection which cannot be packaged. Exowuclease Activity Associated with Tl Gene 4 Several lines of evidence have indirectly suggested that Tl genes 3.5 and 4 specify an exonuclease activity similar in function to that coded for by the red genes of phage X and the RecE genes of E. co& This prompted the search for exonuclease ac-

204

PUGH

AND

tivity in extracts from cells infected with wild-type and urn23 mutant phage. The major exonucleolytic activity in E. coli cells is the ATP-dependent exonuclease V encoded by the recB and recC genes, whereas the X exonuclease and the exonuclease VIII coded by the RecE pathway are ATP independent. Our experimental protocol was designed to distinguish these two types of activity. Figure 3A shows the pattern of ATPdependent and ATP-independent exonuclease activity after infection of E. wli B with Tl+ phage. The ATP-dependent activity is shut off rapidly after infection, whereas there is a corresponding increase in the ability of cell extracts to degrade DNA exonucleolytically in the absence of added ATP. Since this latter activity was present only at low levels at the time of infection (0-min sample), the rise in activity is presumably a consequence of Tl phage infection. The phage-associated exonuclease level rises rapidly from the start of infection, to reach an optimum in-

RITCHIE

tracellular concentration by about 15 min postinfection. By contrast, this ATP-independent activity rises to a much lesser extent above the uninfected level in extracts from nonpermissive cells infected with Tl urn23 mutant phage (Fig. 3B). Figure 3B also shows that am23 infection has no effect on the rapid and efficient shutoff of the recBC exonuclease. A similar inhibition of exonuclease V activity was observed with extracts from nonpermissive cells infected with amber mutants in Tl genes 1,2, and 3.5, which are also essential for Tl DNA replication. The addition of chloramphenicol before or during infection immediately inhibits this decline in exonuclease V activity. The exonuclease present in Tl+-infected cell extracts showed optimum activity at pH 8.0, had an absolute requirement for divalent cations, and was inhibited by ATP. It was equally active on both E.coli and Tl phage DNA substrates, and was particularly active with sonicated DNA substrates. DISCUSSION

- 1.6

-IL

z x f u

0

I B

m=

-

+ - 1.2 m 'a : 10‘; z u - 1.8 ,I -16 - 14 - 1.2

0

I 5

I I 15 10 TIME Imtnl

' 20

1.0

FIG. 3. Exonuclease activity in Tl-infected cells. Cultures of E. co15 B41 were infected with Tl+ (A) and am23 (B) phage and, at intervals, cell-free extracts were prepared with Brij-lysozyme. Nuclease activity in each extract was assayed in the presence (W) and absence (0) of ATP by the release of TCA-soluble radioactivity from ‘H-labelled E. cdi DNA.

Earlier studies have shown that Tl genes 3.5 and 4 play a central role in Tl phage DNA replication and recombination (Christensen, 1976; Ritchie, et uL, 1980; Ritchie and Joicey, 1980; Pugh and Ritchie, 1934). These two gene functions are essential for the formation of intracellular concatemers, and are responsible for the major pathway in the formation of genetic recombinants. In the present report we show that the capacity for host cells expressing exonuclease VIII to rescue the gene 3.5 and gene 4 defect results from the restoration of concatemer formation. This provides the essential substrate for DNA packaging, without which no infectious progeny phages are produced. The results from Fig. 1 show that the rescue by the RecE pathway is not complete and that the intracellular DNA exists as two populations, concatemeric DNA (presumed to result from exoVII1 activity) and the shorterthan-monomer-length genes that are the sole product of gene 3.5- and gene 4- infections of sup” bacteria. Furthermore, the

RecE FORMATION

OF PHAGE Tl CONCATEMERS

205

concatemers produced in sup’ RecE infec- sist during infection, and that this catations are less abundant and are shorter lyses the DNA degradation observed in than in either Tlf infection or in gene 3.5- gene 3.5- and gene 4- infections (Pugh and or gene 4- infections of sup+ cells. Our data Ritchie, 1934). do not indicate whether this is the result The Tl general recombination system of high levels of exoVIII which act ineffi- (referred to as gm by Ritchie et c& 1980) ciently on Tl DNA or whether the exoVII1 shows a marked similarity to the correis as active as the Tl-induced function but sponding “two-gene” exonuclease systems of phages X (r& and red@)and T4 (genes is present in suboptimal concentrations. Preliminary data, showing that the RecE 46 and 47). The similarity with the T4 repathway also partially restores the defect combination system even includes a recent in genetic recombination observed in gene observation that Tl gene 3.5 and gene 4 3.5- and gene 4- infections of sup0 cells, amber mutants are specifically suppressed also supports the picture of an overlap in by internal suppressor mutations with Tl between the processes of concatemer properties very like those of T4 &LS muformation and recombinant formation. tations (Shah and Berger, 1973; Mickelson The interchangeability of the E. coli and Wiberg, 1981). It would be interesting RecE, X Red and Tl gene 3.5/4 functions to know if the interchangeability of the X, clearly suggested that these Tl genes spec- RecE(ruc), and Tl recombination exonuify an exonuclease activity. The data from cleases also included the T4 genes 46/47 Fig. 3 confirm that Tl induces an exonu- and T7 gene 6 products. clease early after infection, and that this activity is largely absent from extracts of ACKNOWLEDGMENT am23-infected, nonpermissive cells. Although this may not be ultimate proof that This work was supported by Grant GR/A71233 from the Tl-induced exonuclease is the primary the Science and Engineering Research Council. product of genes 3.5 and 4, the weight of supporting information makes this very REFERENCES likely. Mutants in Tl genes 3.5 and 4 share a BARBOUR,S. D., and CLARK,A. J. (1970). Biochemical variety of phenotypic properties. These inand genetic studies of recombination proficiency in clude (a) the premature arrest of phage Escherkhia co& I. Enzymatic activity associated DNA synthesis, (b) inhibition of concatewith recB+ and reef?. Proc. Nat1 Ad Sci. USA 65, 955-961. mer formation, (c) formation of shorterthan-monomer-length daughter molecules, BARBOUR,S. D., NAGAISHI,H., TEMPLIN,A., and CLARK, A. J. (1970). Biochemical and genetic studies of re(d) reduction in level of genetic recombicombination proficiency in Escherichia wZi II. Ret+ nation, and (e) suppression by the RecE revertants caused by indirect suppression of Recrecombination pathway. Nevertheless, they mutations. Proc. Nat1 Ad Sci USA 67,123-135. represent different genetic functions by CHRISTENSEN,J. R. (1976). The Red system of bacstandard genetic and biochemical criteria. teriophage lambda complements the growth of a Amber and ts mutants fall into two distinct bacteriophage Tl gene 4 mutant. J. Viral. 17,713complementation groups, which occupy 717. nonoverlapping but adjacent regions of the FIGURSKI,D., and CHRISTENSEN,J. R. (1974). Functional characterisation of the genes of bacterioTl genetic map. Gene 3.5 codes for a polyphage Tl. Virology 59.397-407. peptide of 20,000 molecular weight (Ritchie GILLEN, J. R., WILLIS, D. K., and CLARK, A. J. (1981). and Joicey, 1983), whereas the primary Genetic analysis of the RecE pathway of genetic gene product of gene 4 has a molecular recombination in Eschxrichia coli K12. J. Backriol, weight of about 40,000 (C. L. A. Paiva and 145.512-532. D. A. Ritchie, unpublished results). KAISER,K., and MURRAY,N. E. (1979). Physical charAlthough the activity of the host RecBC acterisation of the Rat prophage in Escherichia wli enzyme (exonuclease V) is inhibited soon K12. Mel Gen tTen& 175,159-174. after Tl infection, we cannot rule out the KU~HNER,S., NAGAISHI, H., and CLARK, A. J. (1974). possibility that low exonuclease levels perIsolation of exonuclease VIII: The enzyme associated

PUGH AND RITCHIE

206

with &A

indirect suppressor. Proc NatL Ad

Sci USA 71,3593-3597.

MACHAT~IE,L. A., RITCHIE,D. A., THOMAS,C. A., JR., and RICHARDSON,C. C. (1967). Terminal repetition in permuted T2 bacteriophage DNA molecules. J. MoL Bid

23.355-363.

MICHALKE,W. (1967). Erhohte Rekombinations haufigeit an den Enden des Tl-chromosoms. MoL Gm. Genet 99,19-33. MICKELSON,C., and WIBERG,J. S. (1981). Membraneassociated DNaae activity controlled by genes 46 and 47 of bacteriophage T4 D and elevated DNase activity associated with the T4 das mutation. J. ViroL 40, 65-77. MILLER, J. H. (1972). Selection of Thy- strains with trimethoprim. In “Experiments in Molecular Genetics,” pp. 218220. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. PUGH,J. C. (1982). Replication and recombination of bacteriophage Tl DNA. Ph.D. Thesis, University of Liverpool. PUGH,J. C., and RITCHIE,D. A. (1984). The structure of replicating bacteriophage Tl DNA: Comparison between wild-type and DNA-arrest mutant infections. Vi* 135, 189-199. RADDING,C. M., SZPIRER,J., and THOMAS,R. (1967). The structural gene for lambda exonuclease. Prw N&L Acad Ski USA 57,277-283.

RAMSAY,N., and RITCHIE, D. A. (1980). A physical

map of the permuted genome of bacteriophage Tl. MoL Gm Genet. 179,669-675.

RITCHIE, D. A., and JOICEY,D. H. (1978). Formation of concatemeric DNA ae an intermediate in the replication of bacteriophage Tl DNA molecules. J. Gen VimL 41,699-622. RITCHIE,D. A., and JOICEY,D. H. (1986). Identification of some steps in the replication of bacteriophage Tl DNA. virolosy 103, 191-198. RITCHIE, D. A., and JOICEY,D. H. (1983). Correlation of genetic loci and polypeptides specified by bacteriophage Tl. J. Gen vird 64, 1355-1363. RITCHIE, D. A., CHRISTENSEN,J. R., PUGH, J. C., and BOlmQuE,L. W. (1986). Genes of coliphage Tl whose products promote general recombination. Virdogy 105,371-378. SADOWSKI,P. D. (1982). Recombination. In “Nucleases” (S. M. Linn and R. J. Roberts, eds.), pp. 28-33. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. SHAH, D. B., and BERGER,H. (1973). Effect of a genespecific suppressor mutation (das) on DNA synthesis of gene 46-47 mutants of bacteriophage T4D. .I. ViroL 12, 328-333. THOMAS,C. A., JR., and ABELSON,J. (1966). The isolation and characterisation of DNA from bacteriophage. In “Procedures in Nucleic Acid Research” (G. L. Cantoni and D. R. Davies, eds.), pp. 553-561. Harper and Row, New York.