MAPPING THE 5'-TERMINAL NUCLEOTIDES OF THE DNA OF BACTERIOPHAGE X AND RELATED PHAGES* BY RAY WUt AND A. D. KAISER DEPARTMENT OF BIOCHEMISTRY, STANFORD UNIVERSITY SCHOOL OF MEDICINE, PALO ALTO, CALIFORNIA Communicated by Arthur Kornberg, October 28, 1966 The two ends of a molecule of X DNA have a structure which enables them to join together to form circular monomers1-' or linear dimers, trimers, etc.' Many, if not all, temperate bacteriophages have cohesive ends including coliphages 480,4 21, 186, 424, 434,5 and P2.6 The thermal reversibility of cohesion1'7 and the response of the ends of X DNA to Escherichia coli DNA polymerase and III8 suggest that each end has a short single strand, about 15 nucleotides long, which protrudes beyond a long double-stranded interior of about 50,000 nucleotide pairs. The single strand at one end is judged to be complementary to the single strand at the other end because cohesion occurs only between a "left" end and a "right" end.31 Figure 1 shows how a molecule with complementary single strands protruding from its ends can form a circular molecule by base-pairing between the two single strands.

left half right holf' sus sus 31 HO I W) V oCO =

heov4 Strand I i qht strand FIG. L.-Structure of X DNA and formation of circular molecules. The structure of the 5'- termini is represented diagrammatically;88MMSA, SU8SB, and ix are genetic markers;7 other sym- bols are described in the text.

Soon after infection of a sensitive or a lysogenic bacterium, the two ends of the DNA molecule of an infecting phage cohere to each other and the single-strand inter- ruptions are closed by covalent bonds.9' 10 Thus cohesive ends provide a mechanism by which DNA molecules can be joined transiently or permanently. The occur- rence of cohesive ends in many different temperate phages and the possibility that cohesion may provide a general mechanism for the joining of two DNA molecules prompted us to initiate a study of the sequence of bases in cohesive ends. We wish to report here the identification of the 5'-terminal nucleotides of the DNA of phages X, 480, 21, and 186. The method for identifying 5'-termini is that of Richardson.11 It consists in attaching a P"2-phosphoryl residue to the 5'-OH terminus of a native DNA molecule in a reaction catalyzed by polynucleotide ki- nase,12, 13 then hydrolyzing the DNA to yield 5'-mononucleotides, and finally de- termining which nucleotides carry p32. Since DNA isolated from bacteriophage X is biologically active."4 it has also been possible to orient the two 5'-termini with respect to the genetic linkage map. Materials.-The strains of phage employed were X ind- C1 857,15 21gp,1' ,80,17 and 186p.6 Stocks of X and 186 were prepared by thermal induction. Eacherichia coli K12, strain W3350 (X ind- 170 VOL. 57, 1967 BIOCHEMISTRY: WU AND KAISER 171

C, 857) or W3350 (186p), was grown at 350 with aeration in tryptone broth to an optical density of 1.0 (at 590 mjs), induced by heating to 450 for 10 min, then incubated at 370 with aeration until maximal lysis occurred as indicated by a minimum optical density. Stocks of 21 were pre- pared by ultraviolet light induction of C600 (21gp). Stocks of 480 were prepared by infection of W3350 grown in tryptone broth. Phage were purified and the DNA extracted according to the procedure of Kaiser and Hogness,'4 except that phenol was saturated with Tris buffer (0.01 M) and adjusted to pH 7.1 before use. At least 80% of the single strands were free of breaks in the X DNA preparations, as judged by zone sedimentation in 0.1 M NaOH and 0.9 M NaCl.18 The preparations of 21, ,80, and 186 DNA were sedimented in a sucrose gradient at neutral pH; more than 75% of the UV-absorbing material sedimented as a single peak of double-stranded DNA. _y-P32-adenosine 5'-triphosphate(pyro) (ATP) was prepared according to the procedure of Glynn and Chappell.19 Polynucleotide kinase was purified according to the procedure of Richardson,'3 modified by adding ATP (to 0.1 mM) to the solutions used after the ammonium sulfate step to stabilize the and by an additional O-(diethylaminoethyl) (DEAE)-cellulose column step to remove excess nucleic acid. The latter was carried out by diluting the ammonium sulfate fraction with 4 vol of water, and applying the solution to a 6 cm2 X 12 cm DEAE-cellulose column. Most of the kinase activity (60-80%) eluted with 50 ml of 0.01 M potassium phosphate, pH 7.5, 0.01 M 2-mercaptoethanol. The enzyme was concentrated by precipitation with 2.9 gm of ammonium sulfate per 10 ml of eluate, dissolved in 10 ml of the elution buffer, diluted with 70 ml of 0.005 M 2-mercaptoethanol, and processed on DEAE and phosphocellulose columns as described by Richardson." The phosphocellulose fractions were concentrated by precipitation with 3.7 gm of ammonium sulfate per 10 ml of eluate, and the precipitate was dissolved in 2 ml (per 10 ml of eluate) of 0.02 M Tris buffer (pH 7.5) containing 0.01 M 2-mercaptoethanol. The product had a specific activity of 30-50,000 units/mg and contained no activities detectable by the following tests. Thirty units of kinase were incubated for 60 min with X DNA under the condi- tions described below for the phosphorylation of DNA. The pattern obtained when the product was sedimented as a band in alkaline solution was the same as that given by the DNA before exposure to the kinase. P32-labeled E. coli DNA (20 miumoles, 8 X 106 cpm/,smole) was incu- bated with 5 units of kinase for 30 min, no increase (less than 0.05%) in acid-soluble P32 was found. Venom (Worthington Biochemical Corp.) was further purified to remove small amounts of 5'- activity.20 Pancreatic DNase and E. coli alkaline (chromatographically purified) were products of the Worthington Biochemical Company. Methods.-Phosphatase treatment of DNA: The incubation mixture contained 0.1 M Tris buffer (pH 8.0), 0.2-0.8 mM DNA (expressed as nucleotide equivalents), 1mM ethylenediamine- tetraacetate (EDTA) (to inhibit contaminating in the phosphatase preparation), and 10-30 ug of /ml. After incubation for 60 min at 370, the reaction mixture was extracted twice with buffer-saturated phenol and dialyzed for 3 days (with two changes) against 100 vol of Tris buffer (0.01 M, pH 8.0) containing 1 mM EDTA. This treatment intro- duced no single-strand breaks detectable by band sedimentation in alkali. Phosphorylation of DNA: The reaction mixture contained 0.07 M Tris buffer, pH 7.6, 12 mM MgCl2, 10 mM 2-mercaptoethanol, 0.12-0.3 mM DNA (or phosphatase-treated DNA), 0.03 mM y-P32-ATP (specific activity 1-4 X 109 cpm/lAmole), and 30 units of polynucleotide kinase/ml added in two equal portions at 0 and 30 min. After incubation at 370 for 60 min, EDTA was added to 20 mM and the mixture was chilled and passed through a 1 X 35-cm column of Sephadex G50 (100-300-,u particle size) to remove the unreacted -y-P31-ATP, and then sedimented in a gradient of 5-20% sucrose containing 1 M NaCl, 0.01 M Tris buffer, pH 7.1, and 0.001 M EDTA at 60,000 X g for 12 hr. In some experiments the mixture was layered directly on a sucrose gradient and centrifuged. Intact DNA molecules were collected in fractions 5 through 8 out of a total of 17 fractions. Of the input DNA, 80% was usually recovered as intact molecules and the ratio of P32 to A260 was constant in these fractions. The p32 content of the phosphorylated DNA isolated in the peak fractions was determined after the addition of salmon sperm DNA as carrier and precipitation of the DNA by 0.5 N perchloric acid. The precipitated DNA was dissolved in 2 ml of 0.05 N NaOH and reprecipitated with perchloric acid; the precipitate was then col- 172 BIOCHEMISTRY: WU AND KAISER PROC. N. A. S. lected on a glass filter, washed, dried, and the radioactivity determined. The specific radioactivity of the phosphorylated DNA was calculated from the optical density of the DNA of the sucrose fractions and its content of acid-precipitable p32, taking the molar absorbance at 260 m/A as 6,600. 14 Hydrolysis of purified DNA to nucleoside-5'-monophosphates: Phosphorylated DNA was pre- cipitated with 10% trichloroacetic acid and sedimented. The precipitated DNA was washed twice with 3-ml portions of cold 80% ethanol and resuspended in 0.1 ml of water. The DNA was then digested by the addition of 0.05 ml of 0.1 M Tris acetate (pH 8.0), 0.01 ml of 0.1 M magnesium acetate, and 0.01 ml of pancreatic deoxyribonuclease (2 mg/ml), and by incubation at 370 for 1 hr. The reaction mixture was centrifuged and the clear supernatant solution (con- taining partially digested DNA) was transferred to another test tube. It was further digested after the addition of 0.03 ml of 0.1 M glycine acetate (pH 9.0) and 0.01 ml of venom phospho- diesterase (1500 units/ml), and by incubation at 37° for 45 min. Essentially all of the radio- activity was converted to mononucleotides as judged by chromatography. The reaction mixture was chilled to00 and aliquots were applied to 589 orange-ribbon filter paper (Schleicher and Schfll, Keene, N.H.), together with 0.05-0.1 /Lmole each of the four deoxyribonucleoside-5'-monophos- phates as carriers and markers. Chromatography was carried out for 8-12 hr in the solvent system of Markham and Smith2l modified as follows: saturated ammonium sulfate, 80 parts; 1 M sodium acetate, 18 parts; isopropanol, 2 parts. In some experiments the chromatographic analysis was confirmed by paper electrophoresis, 100 min at 2400 volts in 0.05 M sodium citrate, pH 3.5. DNA infectivity assays were performed as described previously.7 Results.-Extent of phosphorylation of DNA from phages X, 21, q80, and 186: Polynucleotide kinase catalyzes a specific transfer of the y phosphate of ATP to the 5'-hydroxyl termini of polynucleotides.12' 13 Using 'y-P32-labeled ATP as the donor, ap32 residue can be affixed to the 5'-terminus of each DNA strand. The DNA isolated from phages X, 21, 480, or 186 accepted little or no phosphate unless it was first exposed to bacterial alkaline phosphatase (Table 1). The fixation ofp32 catalyzed by polynucleotide kinase was in every case increased more than ten- fold by prior treatment of the DNA with phosphatase. Thus all four types of DNA probably carry 5'-phosphoryl groups.22 After treatment with alkaline phosphatase, subsequent phosphorylation of each of the phage DNA's reached a limit. Increasing the level of enzyme from 30 to 45 units, or incubation for 90 rather than 30 minutes did not increase the extent of phosphorylation more than 5 per cent. The limit of phosphorylation was in each case close to that expected from the DNA molecular weights determined by physical TABLE 1 EXTENT OF PHOSPHORYLATION DNA Phosphatase Polynucleotide , -Moles/106 Moles DNA-P- source pretreatment kinase p52 incorporated Expected - - 0.10 _ + 0.23 + - 0.12 + + 1.98, 1.94 2.0-2.1 ,80 - - 0.04 - + 0.10 + + 1.98 2.1-2.2 21 - + 0.14 + + 2.48 2.5 186 - + 0.06 + + 1.70

DNA was exposed to -y-P32-ATP and 30 units of polynucleotide kinase, then sedimented in a sucrose gradient at neutral pH, andp32 sedimenting with whole molecules of X DNA was measured. The expectedP32 incorporation was calculated from molecular weight values reported in the literature: X, 33 105,29X and 31 X 106;30 0 80, 0.95 X;4 21, 26 106.5X VOL. 57, 1967 BIOCHEMISTRY: WU AND KAISER 173 methods and listed in the last column of Table 1; therefore this is a valid method for the specific labeling of the 5'-termini of X DNA. When 5'-p32 X DNAwas treatedwith bacterial alkaline phosphatase, 90 per cent of the terminal phosphate was removed as shown by lines 1, 2, and 3 of Table 2. The remaining 10 per cent seems to be protected against hydrolysis. The resistance of 5'-P-termini, which must have been sensitive at one time since they accepted a p32- phosphoryl residue, suggested the possibility that some of the 5'-P32-DNA had cohered and that termini included in cohered ends are protected against alkaline phosphatase. This possibility was tested by an experiment reported in the lower part of Table 2. Labeled 5Ip32 X DNA was caused to cohere by heating it to 750C and allowing it to cool slowly through the temperature range of the cohesive helix- coil transition. Lines 4, 5, and 6 of the table show that 80-90 per cent of the mole- cules had cohered as judged by loss of infectivity and that the 5'-P32-termini of these molecules are resistant to hydrolysis by alkaline phosphatase since only 25 per cent of the radioactivity was made acid-soluble. The 5'-P32-termini of these molecules became sensitive to phosphatase when the cohered ends were separated (lines 7 and 8). Infectivity of X DNA with 5'-OH or 5'-P-termini: Since the two ends of a molecule of X DNA become covalently joined under certain conditions of infection,', 10 experi- ments were designed to determine whether the infectivity of X DNA for helper-in- fected bacteria would be affected by the presence or absence of phosphate at the 5'- termini. The specific infectivities of (1) native X DNA, (2) native DNA treated with 15 ,ug/ml alkaline phosphatase for 60 minutes at 370, (3) 5'-P32-DNA, and (4) 5'-P32-DNA treated with alkaline phosphatase were measured. The second phos- phatase treatment removed 90 per cent of the p32. The specific infectivities of the four DNA samples ranged between 1.1 and 1.3 X 109 plaques/100 msmoles of DNA. Surprisingly, DNA infectivity is not affected by the presence or absence of a 5'-P group, which may mean that helper-infected E. coli K12 contain a DNA-phospho- rylating or DNA-dephosphorylating system. Identification of 5'-terminal nucleotides: The 5'-P32-DNA after purification by zone sedimentation was hydrolyzed to give nucleoside-5'-monophosphates by the successive action of pancreatic deoxyribonuclease and snake venom phosphodi- . Chromatographic and electrophoretic analysis of the hydrolyzed products

TABLE 2 SENsITIvITr OF p32 AT THE 5'-TERMINI OF LINEAR AND COHERED X DNA TO HYDROLYSIS BY ALKALINE PHOSPHATASE Cohered molecules 5'-Terminal P32 Treatment (%) (cpm) 1. None 0 92 2. Alkaline phosphatase, 6 lug 0 8 3. Alkaline phosphatase, 12 mug 0 8 4. Heat to 750, cool slowly 81 80 5. As #4, then phosphatase, 6 ,ug 88 63 6. As #4, then phosphatase, 12 ,ug 87 59 7. As #4, then heat to 750 again, cool quickly 0 80 8. As #7, then phosphatase, 6 1g 0 8 Since cohered molecules have little or no biological activity toward helper-infected bacteria,7 and since cohered ends separate when the solution is heated to 750,1 the number of cohered molecules is measured by the difference in biological activity before and after heating to 75°. If X is the activity before heating and Y the activity after heating to 75° and rapid cooling, the per cent of cohered molecules - 100 (Y - X)/Y. The activities were measured before phosphatase treatment. The conditions of phosphatase treat- ment are given in Methods and the 5'-terminal P32 was measured as acid-insoluble radioactivity. 174 BIOCHEMISTRY: WU AND KAISER PROC. N. A. S. revealed that (Table 3) the P32 in each of the four phage DNA's tested was almost equally divided between deoxyadenylic acid and deoxyguanylic acid with much smaller amounts in deoxythymidylic and deoxycytidylic acid. Therefore all four types of DNA molecules have deoxyadenylate and deoxyguanylate at their 5'- termini. The 5'-termninal nucleotides of the right and left ends of X DNA: One interpretation of the 5'-terminal analysis of X, 21, 480, and 186 is that a deoxyadenylate residue occupies one 5'-terminus, and a deoxyguanylate residue the other 5'-terminus, on every molecule. Alternatively both ends could be terminated by deoxyadenylate (A) on some molecules and by deoxyguanylate (G) on others. To distinguish be- tween these two possibilities a 5'-terminal analysis was performed on separated half- molecules of X DNA produced by breaking whole molecules with hydrodynamic shear. The two half-size molecules have slightly different base compositions, and thus different buoyant densities in neutral CsCl, permitting their separation by equilibrium density gradient sedimentation.23 Half-molecules generated by hydro- dynamic shearing of 5'-p32 X DNA yielded a single but very broad peak (Fig. 2). 0.4

0.3 rInrectivit 5'-Nuclcotide t Froction Sus9 G A Soo : 9 97 393 7 0 73 27 69 31 0 1 27 73 19 81 2 1189 6 94 L n

0.1

9 1O 1I 12 ~~~~~~~0 0 5 10 14- Fraction numbar FIG. 2.-5'-Terminal nucleotides of left and right half-molecules of X DNA. Curve shows distribution of UV-absorbing material in neutral CsCl gradient23 given by 5'_p32 X DNA stirred at 4500 rpm for 30 min. More than 99% of the molecules had been broken, as judged by loss of linked SuSASuSBiX infectivity. Inset shows distribution of infectivity and 5'-terminal nucleotide as per cent of total infectivity or total radioactivity recovered. Total radioactivities based on 400 cts in low-background (1.5 cpm) counter were: frac 9, 26 cpm; 10, 58; 11, 62; 12, 23. Analysis of four fractions obtained across the peak (inset) showed a satisfactory separation of the two half-molecules, in that 97 per cent of the biological activity in fraction 9 carried the genetic markers sus+SUS+A, located at the left end of the linkage map, while 89 per cent of the biological activity of fraction 12 carries the genetic marker i", located on the right half of the linkage map. The distribution of each of the P32-labeled nucleotides is correlated with the distribution of the genetic markers: the sus+sus+ half is linked to dGMP32 and the is half is linked to dAMP32. The left end of the recombination map therefore corresponds to the end of VrOL. 57, 1967 BIOCHEMISTRY: WU AND KAISER 175

TABLE 3 IDENTIFICATION OF THE 5'-TERMINAL NUCLEOTIDES OF FOUR TEMPERATE PHAGE DNA's Per Cent of Total Radioactivity in DNA- Nucleotides X +80 21 186 dAMP 48 48 50 45 dTMP 3 6 1 7 dGMP 47 44 48 42 dCMP 2 2 1 6 DNA was first treated with alkaline phosphatase and then phosphorylated with 'y-P32ATP in the polynucleotide kinase system (Methods). The phosphorylated DNA was purified, hydro- lyzed, chromatographed on paper, and the radioactivity of the four nucleoside-5'-phosphates was determined (Methods). Radioactivities were measured in a low-background counter (1.5 cpm background) and at least 200 counts were recorded for each sample. Of the total radioactivity put on the chromatography paper at least 95% was recovered in the areas corresponding to the four nucleotides. One hundred per cent radioactivity corresponds to 144 cpm for X DNA, 382 cpm for 080 DNA, 489 cpm for 21 DNA, and 377 for 186 DNA. Two additional experiments with X DNA gave similar results. the molecule having a 5'-G terminus, and the right end of the map corresponds to the 5'-A terminus. The 5'-terminal nucleotide of the two strands of X DNA: Another way of separating the two 5'-termini from each other is to separate the double helix into its constituent single strands. The buoyant densities of the two strands of X DNA in alkaline CsCl (pH 12.0) differ by 0.004 gm cm-3, permitting their partial resolution in a CsCl density gradient.24 The degree of separation obtained with 5 Xp32X DNA is shown in Figure 3. The gradient fractions were divided into five groups, and the distribu- tion of p32 between dAMP and dGMP after hydrolysis of each of the groups is presented in the inset to Figure 3. The densest group, number 1, enriched for the heavier strand, contains three times more dGMP32 than dAMP,32 whereas the least dense group, number 5, enriched for the light strand, contains four times more dAMP32 than dGMP32. The intermediate groups show a transition from one ex- treme to the other. Thus we conclude that the heavy strand has G at its 5'- terminus and the light strand has A. Discussion.-Knowledge of the 5'-terminal nucleotides corresponding to the left and right ends of the genetic map and of the heavy and light strands completely determines the orientation of the X DNA molecule as summarized in Figure 1. The heavy strand has deoxyguanylate at its 5'-terminus, and since the left end of the molecule marked by sus+sus+ has 5'-terminal G, the heavy strand must be oriented with its 5'-terminus on the left and its 3'-terminus on the right. Hogness and Doerfler24 found a pattern of marker inactivation by exonuclease I treatment of the isolated heavy strand which agrees with this assignment. According to the scheme presented in Figure 1, the rules of base-pairing imply that there is a C residue in the proximal part of the right cohesive end and a T residue in the left. Each of the fragments produced when X phage DNA is broken by hydrodynamic shear has been found to carry a unique set of genetic markers.7' 5, 26 Therefore, the genetic markers of each whole molecule are physically ordered in the same linear sequence; by inference each molecule has the same sequence of nucleotides. The work reported here supports this proposition in that, to a confidence level of 5 per cent, all X DNA molecules have deoxyadenylic acid (A) at the 5'-terminus corre- sponding to the right end of the genetic map and deoxyguanylic acid (G) at their left 5'-terminus. The finding that the DNA of 21, 480, and 186 also have only two 5'- terminal nucleotides suggests that these phages also have nonpermuted nucleotide sequences. 176 BIOCHEMISTRY: WU AND KAISER PROC. N. A. S.

L S'A 26141b51 861 79 O 0.4 Refractionated C) C a Cj.5 -0 .0 <0.2-

O.1I 1 2 3 4. 5

0 5 1 0 IS 20 25 30 35 40 Froction number FIG. 3.-5'-Terminal nucleotides of the two strands of X DNA. Curve shows distribution of UV-absorbing material in alkaline CsCl gradient. 5'4_82 X DNA adjusted to pH 12.1 in 0.04 M sodium phosphate, 0.008 M EDTA, CsCl of initial density 1.72 gm cm-3, was centrifuged 36 hr at 136,000 X g.21 Inset shows distribution of radioactivity between deoxyguanylic and deoxy- adenylic acids among five groups of fractions in region of absorbance peaks. Groups 1, 3, and 5 were purified further by sedimentation in an alkaline sucrose gradient before hydrolysis 80% of DNA input was recovered as intact single strands. Total radioactivities based on more than 400 ots in a low-background (1.5 cpm) counter were: group 1, 33 cpm; 2, 29; 3, 36; 4, 38; 5, 15.

Bacteriophages X, 480, and 21 are closely related to each other; genetic recombi- nation has been found between 21 and X'6 and between 480 and X17, 27 Moreover, the cohesive ends of X, 21, and 080 must be similar, if not identical, because 4)80 DNA and 21 DNA will form mixed dimers with X DNA.4 6 It is therefore not unexpected that the 5'-terminal nucleotides of 480 and 21 should be the same as those of X. Phage 186, on the other hand, is not closely related to X; these phages differ antigenically and in gross morphology, and no recombinants have yet been found in crosses between X and 186. Phage 186 DNA has cohesive ends because it will form dimers, but it will not form mixed dimers with X DNA. The 186 cohesive ends are, nevertheless, structurally similar to those of X because they are biologically inacti- vated by DNA polymerase-catalyzed repair and reactivated by exonuclease III- catalyzed hydrolysis8' 28 and the 5'-termini are A and G. Perhaps the ends of 186 DNA differ only in the details of their base sequence from those of X. Summary.-The DNA molecules of four temperate coliphages X, 480, 21, and 186 have deoxyadenylate and deoxyguanylate at the 5'-termini of the two constituent strands. In X DNA, deoxyguanylate is at the left end of the double-stranded mole- cule corresponding to the genetic marker SUSA, and deoxyadenylate is 5'-terminal at the other end. The more dense of the two strands of X DNA in alkaline CsCl has 5'-terminal deoxyguanylate and therefore is oriented with its 5'-terminus at the SUSA end (Fig. 1). VOL. 57, 1967 BIOCHEMISTRY: WU AND KAISER 177

Mrs. Plana Barrand provided very able assistance in these experiments. Abbreviations used: EDTA, ethylenediaminetetraacetate; DEAE, diethylaminoethyl; -y- P32-ATP, adenosine triphosphate labeled with p32 in the terminal phosphate; 5'-P'2-DNA, 5'_p32 (phosphoryl) DNA; A, deoxyadenylate; G, deoxyguanylate; C, deoxycytidylate; T, deoxy- thymidylate. * This work was supported by research grant AI 04509 from the National Institutes of Health and grant CA 05706 from the National Cancer Institute. t On leave from the Public Health Research Institute of the City of New York. Present address: Department of Biochemistry and Molecular Biology, Cornell University. 1 Hershey, A. D., E. Burgi, and L. Ingraham, these PROCEEDINGS, 49, 748 (1963). 21Ris, H., and B. L. Chandler, in Cold Spring Harbor Symposia on Quantitative Biology, vol. 28 (1963), p. 1. 3 MacHattie, L., and C. A. Thomas, Science, 144, 1142 (1964). 4Yamagishi, H., K. Nakamura, and H. Ozeki, Biochem. Biophys. Res. Commun., 20, 727 (1965). 5 Baldwin, R. L., P. Barrand, A. Fritsch, D. A. Goldthwait, and F. Jacob, J. Mol. Biol., 17, 343 (1966). 6 Mandell, M., personal communication. 7Kaiser, A. D., and R. Inman, J. Mol. Biol., 13, 78 (1965). 8 Strack, H., and A. D. Kaiser, J. Mol. Biol., 12, 36 (1965). 9 Young, E. T., and R. L. Sinsheimer, J. Mol. Biol., 10, 562 (1964). 10Bode, V. C., and A. D. Kaiser, J. Mol. Biol., 14, 399 (1965). " Richardson, C. C., J. Mol. Biol., 15, 49 (1966). 12 Novogrodsky, A., and J. Hurwitz, J. Biol. Chem., 241, 2923 (1966). 13 Richardson, C. C., these PROCEEDINGS, 54, 158 (1965). 4 Kaiser, A. D. and D. S. Hogness, J. Mol. Biol., 2, 392 (1960). 16 Jacob, F., and A. Campbell, Compt. Rend., 248, 3219 (1959). 6 Kulke, M., Ph.D. thesis, Stanford University, Stanford, California, 1964. 17 Franklin, N. C., W. F. Dove, and C. Yanofsky, Biochem. Biophys. Res. Commun., 18, 910 (1965). 8 Studier, F. W., J. Mol. Biol., 11, 373 (1965). 19 Glynn, I. M., and J. B. Chappell, Biochem. J., 90, 147 (1964). 20 Sinsheimer, R. L., and J. F. Koerner, J. Biol. Chem., 198, 293 (1952). 21 Markham, R., and J. D. Smith, Biochem. J., 52 552 (1952). 22 C. Richardson and B. Weiss (J. Gen. Physiol., 49, 81 (1966)), have reported that another strain of X, propagated in another host, bears 5'-OH groups. 23 Hershey, A. D., E. Burgi, and C. J. Davern, Biochem. Biophys. Res. Commun., 18, 675 (1965). 24 Hogness, D. S., W. Doerfler, J. B. Egan, and L. W. Black, in Cold Spring Harbor Symposia on Quantitative Biology, vol. 31 (1966), in press. 25 Radding, C. M., and A. D. Kaiser, J. Mol. Biol., 7, 225 (1963). 26 Hogness, D. S., and J. R. Simmons, J. Mol. Biol., 9, 411 (1964). 27 Signer, E. R., Virology, 22, 650 (1964). 28 Kaiser, A. D., unpublished. 29 Caro, L. G., Virology, 25, 226 (1965). 80 Burgi, E., and A. D. Hershey, Biophys. J., 3, 309 (1963). 31 Hershey, A. D., and E. Burgi, these PROCEEDINGS, 53, 325 (1965). 32 Khorana, H. G., in The Nucleic Acids, ed. H. Chargaff and J. Davidson (New York: Aca- demic Press, 1960), vol. 3, p. 105.