ENZYMATIC SYNTHESIS OF DNA, XXIII. SYNTHESlS OF CIRCULAR REPLICATIVE FORM OF PHAGE 4x174 DNA* BY MEHRAN GOULIANt AND ARTHUR KORNBERG DEPARTMENT OF BIOCHEMISTRY, STANFORD UNIVERSITY SCHOOL OF MEDICINE, PAL0 ALTO, CALIFORNIA Communicated August 24, 1967 Replication of a helical DNA template by Escherichia coli DNA polymerase yields a macromolecule with the chemical characteristics expected of a complemen- tary copy of the template.' However, the product as seen in the electron microscope is a branched rather than a simple linear fiber,2 renatures very readily after a denaturing treatmentj2 and shows no biologic activity unequivocally dis- tinguishable from that in the associated tem~late.~ The physical anomalies, which can be ascribed to a failure to replicate both strands of a helical template simul- taneously from one end, were absent when single-stranded circular DNA from phage M13 or 4x174 was the template.' But a test of biologic activity of the product strand was still not practicable inasmuch as the product strand was linear and only circular molecules are known to be infective? With the discovery of enzymes that join properly aligned ends of a DNA mole- cule,'j-l0 the way was open to see whether DNA polymerase in conjunction with a DNA-joining enzyme can synthesize circular 4x174 DNA and whether such syn- thetic molecules are infective. This report describes the synthesis, isolation, and characterization of a fully covalent duplex circular product from 4x174 DNA tem- plates. A succeeding report'l will describe (1) the isolation of the synthesized complementary circles from these duplexes, (2) the ability of these single-stranded circles to serve as templates for the production of completely synthetic duplex circles, and (3) the capacity of enzymatically synthesized single and duplex circular forms to produce phage particles when mixed with spheroplasts of E. coli. Ezperimeniul Procedure.-MateriaIs: Unlabeled deoxynucleoside triphosphates were purchd from Pabst and puri6ed by chromatography on DEAESephadex (A-25) using a triethylammonium bicarbonate gradient. a-P'Z-dCTP and a-PJ2-dATP were obtained from Volk, HJ-methylthymi- dine from New England Nuclear, DPN from Sigma, and CsCl (optical grade) from Harshaw. dmTP was prepared by bromination and deamination of dCTP.18 Ethidium bromide was a gift from Boots Pure Drug Co., Nottingham, England. E. coli DNA polymerase was purified by the method of Jovin el al.13 and E. coli polynucleotide-joining enzyme by the method of Olivera and Lehman.' Pancreatic DNase was the Worthington 1 X recrystallized product. Boiled extract of E. coli was the supertiatant fluid obtained by centrifuging a sonic extract of E. coli strain 1100 (deficient in endonuclease I, obtained from Prof. H. Hoffman-Berling) which had been heated at 100' for 5 niin as described by Olivera and Lehman.8 H'QX174 DNA RF-form I (750 cpm/mpmole) was a gift from Prof. It. L. Sinsheimer; RF- form I1 was generously provided by Dr. P. Pouwels. Methods: Sucrose gradient sedimentations were carried out in 4-ml linear 520% sucrose gradients, in buffers indicated in figure legends and using an IEC B-60 preparative ultracentrifuge with an SB-405 rotor. Fractions were collected and precipitated on filter paper disks with 10% trichloroacetic acid, washed with 0.01 N HCl and then with ethanol, dried, and counted in a toluene-base scintillator solution. Standards were processed in the same way. H"X174 DNA WBS prepared from the lysis-defective amber mutant am 3 using a procedure based on that of Yinsheimer." The thymilie-requiring, noiipermisive host, E. coli 15 THU-, wa? grown in 2 1 of TPG 3A medium containing supplements of L-histidine (20 pglml), uracil (10 ug/ml), and thymidine (4 pg/ml). At the time of infection (4 X 1W bacteria/ml, multiplicity 1723 1724 UIOCHE'BI ISTI< 1': COULI. I S .I .YD KORSUBIZG PROC. N. A. S. of 3), H'-thymidilie (4 mc, 0.14 pg) was added. After the lysis procedure,14 the phage were purified first by CsCl equilibrium density gradient centrifugation and then by velocity sedimenta- tion in a sucrose gradient. The DNA, extracted from the purified phage with phenol, had a specific activity of 13,000 cpm/mpmole based on absorbancy at 260 mp in 0.01 M Tris-HC1, pH 8.0, and 1 mM EDTA and a molar absorbancy of 9,5Oc).l5 Molarity of DNA refers to concentration of nucleotide residues. Electron microscopy was rarried out with the protein film technique of Kleinschmidt el~l.'~aas modified by Inman el d.17 The hypophase was water or 0.1 M ammonium acetate, as indicated. Sterile solutions and vessels were used wherever possible, to minimize nuclease contaminat ion. Opemtioiih iiivolving BIT-containing DNA were carried out in subdued light. Results.-Replication os plraqe (+) circles and isolation oj duplex circular product: When H3+X174 DNA was the template for replication by E. coli DNA poly- merase in the presence of P3*-dCTP and three unlabeled deoxynucleoside triphos- phates, E. coli polynucleotide-joining enzyme, DPN, and a boiled extract of E. coZi, covalently closed duplex circles were formed. The reaction was carried out at 25". The amount of nucleotide incorporation proceeded to approximately one replication of the input DNA. The products were fractionated by equilibrium sedimentation in CsCl in the presence of ethidium bromide.I8 More ethidium bromide is bound to incomplete (or nicked) circles and linear molecules than to covalent duplex circles; for this reason the latter have a higher buoyant density.18 As shown in Figure 1, about half of the phage DNA (51%,) was found in the heavier band (partially synthetic duplex circles),lg in which there was an equimolar ratio of product to template across the peak; in other experiments 38 to 57 per cent of the phage DNA was in the heavy band. The excess of product over template in the lighter band may be due to priming by polynucleotides in the boiled extract or to synthesis beyond one replication' in circular molecules in which joining of the replica failed to take place. Fraction number FIG. 1.-Fractionation of duplex circles by CsCl-ethidium bromide equilibrium sedimentation. The in- cubation mixture (0.50 ml) con- tained 0.18 mM HJ+X174 am 3 DNA, 0.45 mM each of dCTP, dTTP, dGTP, and a-P**-dATP (60,000 cpm/mpmole), E. coli DNA polymerase (1,600 iinits/ml), E. coli joining enzyme (2 units/ml), 8 pM DPN, E. coli boiled extract (40 fil/ml), 5 mM MgC12, 20 mM potassium phosphate buffer, 1 mM 8-merca toethanof*i;d albumin (40 ,/rrf). After incuba- tion for 180 min at 25O, the mixture was adjusted to 3 ml with 10 mM Triscitrate buffer, pH 7.6, contain- ing 10 mM EDTA, 300 pg of ethi- dium bromide, and 2.25 gm of CsCl. The resulting mixture (p = 1.56) was centrifuged in the IEC B-60pre- parative ultracentrifuge using the SB 405 rotor at 45,000 rpm for 40 hr at 'LO". An aliquot. of 1 pl from each fract.ioii was precipitat.ec1 011 3 MM paper disks for counting. In this and succeeding figures, ordinate value3 represent total moles of nil- cleotide per fraction. VOL. 58, 1967 BIOCHEMISTRY: GOULIAA' AND KORNBERG 1725 Replication was also carried out with dTTP replaced by the bromouracil analogue, mTP. The CsC1-ethidi- um bromide equilibrium pattern of the products (Fig. 2) was similar to that in Figure 1. The BU-labeled duplex circles in the heavy band were sepa- rated and used as starting material for isolation of (-) circles as described in a later report." These duplex circles were also used for electron microscopic examination (see below). Velocity sedimentation of partially synthetic duplex circles: An alkaline sucrose gradient analysis of the mole- cules in the heavier band of the CsC1- ethidium bromide equilibrium sedimen- tation provided further evidence that these molecules were duplex covalent circles (Fig. 3). They sedimented in an alkaline sucrose gradient coincident with the natural form of dX174 du- plex covalent twisted circles (RF-form I)" at a rate approximately three times that of phage DKA or of RF- form I1 (which is nicked and sepa- rates into single strands at alkaline PHI. Sedimentation at neutral pH indi- cates the extent of supercoiling (twist- m P m x Ei - 9 1 1 FIG. 2.-Fractionation of duplex circles, con- taining FU in the synthetic (-) strand, by CsCl-ethidium bromide equilibrium sedimenta- tion. The conditions for incubation and frac- tionation by density gradient centrifugation were the same as for Fig. 1 except that the a-Pa* label was in dCTP (50,000 cpm/mpmole) instead of dATP and dmTP replaced dTTP. ~ ing) in a duplex molecule. Supercoiled duplex molecules are more compact and sedi- ment more rapidly than duplex circles which are relaxed.*' If the partially synthetic duplex circles were supercoiled, their sedimentation would resemble that of RF- form I; with little or no twist they would sediment more like RF-form I1 which is nicked and relieved of its twist. At neutral pH and low ionic strength the partially synthetic duplex circles sedimented less rapidly than RF-form I and had the same sedimentation coefficient as the small amount of form I1 contaminating the prep- aration of RF-form I (Fig. 4). The partially synthetic duplex circles sedimented at a rate of 0.76 times that of the natural RF-form I and this may be compared to the value of 0.71 determined for natural form II.22 At neutral pH and high ionic strength (1 M NaCl), the partially synthetic duplex circles also sedimented more slowly than natural RF-form I; the rate relative to form 1 was 0.85, whereas the value determined for forin I1 was 0.79,21 and again the position coincided with con- tnininutiiig form 11. The eiizymatically synthesized molecules were examined in the electron microscope using water rather than salt solution as the hypophase. All molecular forms were photographed Electron microscopy oj the parliall!j syntlietic duplex circles: 1726 BIOCHEMISTRY: GOULJAN AND KORNBERG PROC. N. A. S. I RF Form I ' 4 Phoqa DNA I ' marker morhur FIG. 3.-Alkaline sucrose gradient sedimentation of purified partially syn- thetic duplex circular molecules. Frac- tions 17 and 18 from the gradient shown in Fig. 1 were combined and dialyzed, first against 1 M NaC1, 10 mM Tris- HC1, pH 8, 1 mM EDTA and then against the same buffer witkout NaC1. A 3 &aliquot of the dialyzed fraction was centrifuged in a sucrose gradient in 0.2 M NaOH, 0.8 M NaCl, 1 mM EDTA at 60,000 rpm for 75 min at 10". Additional tubes in the same cen- trifugation contained samples of Ha- qjX174 DNA and Ha-natural FtF-form I, both alone and mixed with the dia- lyzed CsCl fraction, to serve as markers for their relative positions in the gradient. Fraction number Fraction number FIG. 4.-Neutral, low-salt, sucrose gradient sedimentation Of p urified . partially synthetic dup ex circular molecules. The dialyzed and purified prepara- tion of duplex circles was the same as in Fig. 3. Sucrose eradient sedimentation was car- Red out in 9 mM NaCI, 1 mM Tris-HC1, pH 8,O.l mM EDTA, at 60.000 rDm for 210 min at 10". The s'ame r'eference markers were used as in Fig. 3, and were placed in the tube containing the sam- ple and also run separately. For clarity onl the Pa* of the syn- thetic duprex circles and H* of the natural RF are shown; the Ha label of the synthetic duplex circles coincided with the Pas as in Fig. 3. The data given here are from two separate gradients; the results were iden- tical when synthetic and natural RF were run together in the same gradient. serially without selection, and of these, 75 per cent were simple circles (Table 1, Fig. 5). The average contour length was 1.97 p (Table 1) with a distribution skewed toward the shorter molecules. A preparation of natural RF-form I1 was examined for contour length and showed approxiinately the same fraction of simple circles with an average length of 2.07 p. When the same natural RF preparation was spread on 0.1 M ammonium acetate as hypophase, the average length (1.55 p) was in better agree- ment with published values of 1.64 * 0.1122 and 1.77 f 0.0g23 in which the salt hypophase was used. This capacity of salt in the hypophase to shorten the contour Highly twisted forms like those of RF-form IZ3* 24 were not present. I'oL. 38, 1967 BIOCHEMISTRY: GOULIAN AND KORNBERG 1727 TABLE 1 DIS'I-RIBUTION AND CONTOUR LENGTHS OF CIRCULAR FORMS IN RF PREPARATIONS --Contour Length- Circles .4verage RF Hypo- Total forms Linear with Circles and S.D. preparation phase photographed forms Circles* branches measured G) Partially synthetic's Water 301 11 225 12 203 1.97 f 0.12 Natural Water 73 9 51 3 46t 2.07 f 0.07 Natural Salt 1 56 6 41 2 41 1.85f0.14 * Includes only those circ!es whose circumference could be traced clearly: aggregated, superimposed, or t Doea not include three large circles (see text). $0.1 M ammonium acetate. partially twisted forma constituted the remainder of the structures not classified in this table. length has been clearly demonstrated by Inman.*j A branchlike structure which varied in length from 0.05 to 0.36 p was found on a few of the molecules, synthetic as well as natural. Another anomaly appeared among the natural RF circles. Of 97 circles examined, 3 had lengths of 4.20, 4.25, and 4.42 p, values just twice the average size; no such form was seen among 237 partially synthetic duplex circles. Op- timal yields of the covalently closed duplex circles required all the components of the complete system (Table 2). There was an absolute requirement for polym- erase, triphosphates, and phage DNA, whereas DPN, the cofactor for the joining enzyrne,26- Join- ing enzyme was active at a level expected of its activity in completing X circles.28 The rather high conversion in the absence of added joining enzyme suggests that the purified polymerase may be contaminated with polynucleotide-joining enzyme; however, adequate assays have not yet been performed to teat this pos~ibility.~~ The function of the boiled extract will be discussed below. . Discussion.-Replication of the circular phage DNA by DNA polymerase alone In$mce of various factors on replication and the synthesis of duplex circles: had the least influence due to its presence in the boiled extract. had been judged to be complete chiefly on the basis of observations in the electron microscope of duplex circles with contour lengths of 2 p.' However,, even the ap- pearance of a full helical circle would not have disclosed a discontinuity in the mole- cule of as much as 0.5 per cent, represent- ing a gap of 30 nucleotides. Since the polynucleotide-joining enzyme requires that the 3'-hydroxyl and 5'-phosphate termini be closely aligned for joining, the success of the conjoint effort of this en- zyme with polymerase in completing a fully covalent duplex circle (RF) makes it more likely that replication of the phage circle by polymerase does go to completion. Also remarkable was the efficiency with which the replicas were joined; the values for the amount of RF synthesized indicate t8hat near 50 per cent of the phage DNA was converted. FIG. 5.-Electron micrographs of partially synthetic duplex circles. The DNA sample was obtained from the peak tubes of the heavy band in Fig. 2 and dialyzed (see legend to Fig. 3). For details of electron microscopy, see the text. 1728 BIOCHEMISTRY: GOULIAN AND KORNBERG PROC. N. A. s. TABLE 2 Although the experiments described in this report have been limited to the repli- cation of 6x174 by E. coli polymerase, Minus polymerase <0.2 studies with another pha.ge DNA and with " dTTP <0.2 another DNA polymerase have yielded '' phage DNA <0.2 " DPN 26 comparable results. Phage M13 DNA, 4.5 which resembles 6x174 DNA in structure " joining enzyme " boiled extract 1.9 and in size,', was also converted by the described for Fig. 1. After incubation, each mix- E. coli polymerase-polynucleotide-joining in a sucrose gradient in 0.2 M NaOH, 0.8 M NaCI. enzyme system to a duplex Circular mole gradient was then divided into 10 fractions and the cule. The product, synthesized under was pooled and is expressed as yo of total H' in the conditions described for 4x174 DNA (Fig. l), had the same characteristics as partially synthetic 6x174 RF in equilibrium sedimentation in the presence of ethi- dium bromide, and in velocity sedimentation at alkaline and neutral pH. A duplex circular molecule was also produced by the combined action of E. coli and phage T4 DNA polymerases. A 3'-hydroxyl initiator site, required by the phage polym~rase,~' was provided by annealing small pieces of complementary (-) strands to M13 DNA circles. Such fragments were prepared by partial replication of M13 circles with E. coli polymerase and subsequent fractionation of the products on alkaline sucrose gradients. The M13 circle with fragments annealed to it replicated by the phage polymerase and the complementary circle was closed to form a duplex by joining-enzyme activity (ligase)', lo that contaminated the purified phage T4 polymerase preparat,ion. The electron microscopic appearance of the partially synthetic RF is that of a circle of t.he same size as RF isolated from infected E. coli but with little or no supercoiling. Branchlike structures, which occurred in only 5 per cent of the circular molecules, were no more common than in the natural RF population and may repre- sent a tightly twisted portion of the circle. Although large molecules, twice the average size, were seen in 3 of 97 natural RF molecules examined, no such giant forms were found in 237 synthetic molecules nor were there any midgets, either. The significance of the giant forms of RF, possibly a result of' an intracellular recombinational event between two RF molecules, deserves further investigation. GellertO and Gefter et aL9 have reported that X DNA circles closed in vitro sedi- ment more slowly in neutral sucrose gradients than natural duplex circles, and under some conditions resemble nicked duplex circles. The partially synthetic duplex circles described here have a similar relationship to natural RF-forms I and I1 or 4x174 DNA in sedimentation at neutral pH. We assume that the structures syn- thesized in these studies, as in the case of enzymatically closed X circles, have, under the conditions employed, little or no twist. The extended appearance of the synthe- sized circles in the electron microscope supports this assumption. The mechanisin by which twist is introduced at the time of closure is not yet clear, t the effects ceytration on DNA structure in general,32 and particularly as it applies to duplex cjrcles,6 is well documented. The recent work of Wang33 makes it clear that the ,temperature at the time of closure of the duplex circle has a profound effect. This effect may account to a major extent for the lower amount of twist in the duplex Conversion of ,,IIaEe DNA to RF (%) Complete system 38 The complete system, in 11 pl, was the same as ture was made 60 mhl in EDTA and centrifuged 1 mM EDTA at 60,000 rpm for 60 min at 10'. The H' in the rapidly sedimenting fractions (see Fig. 3) gradient. of salt and tempwature are probably very important. The influenc 9 of salt con- VOL. 58, 1967 BIOCHEMISTRY: GOUUAN AND KORNBERG 1729 circles formed enzymatically at 25" in our studies as compared to those produced in E. coli at 37". An aspect of the RF synthesis which is not properly clarified is the stimulation of the over-all reaction by some component in a boiled extract of E. coli. Ex- ploratory studies concerning the function of boiled extract in this system have shown that (1) the presence of joining enzyme reduced polymerase replication of +X174 DNA to a value of 12 per cent in its absence; addition of boiled extract restored it to a level of 120 per cent; (2) the extract stimulated replication of cir- cular DNA 1.8-fold in the absence of joining enzyme; (3) the latter stimulation wlls even greater when the extract was first treated with pancreatic DNase; and (4) the DNase-treated extract was inhibitory in the replication-joining system at concentrations optimal for the replication reaction alone but retained activity when used at lower levels. The effect of boiled extract therefore appears to be on the replication rather than the joining part of the over-all synthesis. Some of the uncertainties regarding the start of replication of a circular template' suggest that the effect may be in the initiation phase of the reaction. Perhaps a very small piece of DNA in the extract adheres to the phage circle and provides a priming point for replication. The effect may also be related to the inhibitory action of the joining enzyme on replication, which is also unexplained. More details on how replication by E. coli polymerase starts and proceeds on a circular template might contribute to the explanation of the stimulation by boiled extract. The capacity of DNA polymerase and joining enzyme to produce a circular complementary copy of phage DNA has now made it possible to isolate these copies, demonstrate their infectivity, and proceed with the total enzymatic synthesis of infective RF." The implication for studies of mutagenesis becomes clear at once, inasmuch as a variety of base analogues or ribonucleotides can now be incorporated into an infectious molecule. In the same vein, defective polymerases from phages with mutations in the polymerase structural gene can be examined and exploited for errors in their replicative activity. Finally, it should be possible and worth- while to carry out the replication of the circular DNA's from other viruses, such as polyoma, and from cellular organelles, such as mitochondria and chloroplasts. With these duplex circular DNA's as with RF of 4x174," controlled nuclease cleavage and denaturation should release (+) and (-) circles as templates for replication and eventually total synthesis of the duplex structures. Summuq.-Phage 4x174 DNA [(+) circles] were converted by the combined action of E. coli DNA polymerase and polynucleotide-joining enzyme to a duplex circular form (RF). The partially synthetic RF has (1) the sedimentation and dye-binding properties expected of a fully covalent duplex circle, (2) a circular appearance in the electron microscope, and (3) the contour length of natural RF. Partially synthetic RF was also prepared with bromouracil in place of thymine in the (-) circle to facilitate isolation of the (-) circle, to enable the synthesis of a fully syntheticRF, and to demonstrate infectivity of synthetic molecules in studies to be reported. We gratefully acknowledge the coritribution of Miss Maria Schriljs in preparing and examining We also appreciate the helpful advice of Prof. R. L. the specimens for electron microscopy. Siheimer arid Mr. John Newbold in the preparation of H8+X174 am 3 DNA. 1730 BIOCHEMISTRY: GOULIAN AND KORNBERG PROC. N. A. S. Abbreviations: RF, replicative form; DEAE, 0-( diethylaminoethy1)cellulose; dATP, dBUTP, dCTP, and dTTP, deoxynucleoside 5'-triphosphates of adenine, bromouracil, cytosine, and thy- mine; DPN, diphosphopyridine nucleotide; EDTA, ethylenediaminetetracetate. * Supported in part by a grant from the National Institutes of Health, USPHS. t Special fellow, USPHS. Present address: Department of Medicine, University of Chicago. 'Kornberg, A., Enzynzdic Synthesis of DNA, Ciba Lectures in Microbial Biochemistry (New York: John Wiley & Sons, Inc., 1961). Schildkraut, C. L., C. C. Richardson, and A. Kornberg, J. Mol. BWZ., 9, 24 (1964). Lederberg, in Informational Macromolecules (New York: Academic Press, 1963), p. 13. 'Richardson, C. C., C. L. Schildkraut, H. V. Aphian, A. Kornberg, W. Bodmer, and J. Mitra, S., P. Reichard, R. B. Inman, and A. Kornberg, J. Mol. BWZ., 24,429 (1967). 'Fiers, W., and R. L. Sinsheimer, J. Mol. BWl., 5, 424 (1962). 6 Gellert, M., these PROCEEDINGS, 57, 148 (1967). 7 Weiss, B., and C. C. Richardson, these PROCEEDINGS, 57, 1021 (1967). * Olivers, B. M., and I. R. Lehman, these PROCEEDINGS, 57, 1426 (1967). 0 Gefter, M. L., A. Becker, and J. Hurwitz, these PROCEEDINGS, 58, 240 (1967). 1OCozzsrelli, N. R., N. M. Melechen, T. M. Jovin, and A. Kornberg, BWchem. Biophys. Res. 11 Goulian, M., R. L. Sinsheimer, L. L. Bertsch, and A. Kornberg, to be submitted. 12 Bessman, M. J., I. R. Lehman, J. Adler, S. B. Zimmerman, E. S. Simms, and A. Kornberg, ~Jovin, T. M., P. T. Englund, and L. L. Bertsch, in preparation. 14 Sinsheimer, R. L., in Procedures in Nucleic Acid Research, ed. G. L. Cantoni and D. R. Davies 16Sinsheimer, R. L., J. Mol. BWl., 1, 37 (1959). 1' Kleinschmidt, A. K., D. Lang, D. Jacherts, and R. K. Zahn, Bwchim. Biophys. Acta, 61,857 17Inman, R. B., A. Kornberg, and C. L. Schildkraut, J. Mol. Bwl., 11, 285 (1965). 18Radloff, R., W. Bauer, and J. Vinograd, these PROCEEDINGS, 57, 1514 (1967). 1) The term "partially synthetic duplex circle," as used here, refers to a structure in which the Commun., 28, 578 (1967). these PROCEEDINGS, 44,633 (1958). (New York: Harper and Row, 1966), p. 569. (1962). (+) circle is phage DNA and the (- ) circle is enzymatically synthesized. Burton, A., and R. L. Sinsheimer, J. Mol. Biol., 14, 327 (1965). 21Vinograd, J., and J. Lebowitz, J. Gen. Physwl., 49, No. 6, Pt. 2, 103 (1966). "Kiger, J. A., Jr., E. T. Young, 11, and R. L. Sinsheimer, personal communication. *sKleinschmidt, A. K., A. Burton, and R. L. Sinsheimer, Science, 142, 961 (1963). 14 Roth, T. F., and M. Hayashi, Science, 154, 658 (1966). "Olivera, B. M., and I. R. Lehman, these PROCEEDINGS, 57, 1700 (1967). 27Zimmermaq S. B., J. W. Little, C. K. Oshinsky, and M. Gellert, these PROCEEDINQS, 57, 1841 (1967). Assuming one covalent bond closure by the joining enzyme for each complementary circle, a value of 0.04 ppmole min-1 unit-' ww obtained compared to 0.06 for closure of hydrogen-bonded circld and the value of 1 under optimal assay conditions.8 2) The boiled extract which by direct assay8 contained, per ml, less than 0.01 unit of joining Ray, D. S., H.-P. Bscheider, and P. H. Hofschneider, J. Mol. BWZ., 21, 473 (1966); Ray, Inman, R. B., J. Mol. BWZ., 25, 209 (1967). enzyme activity, is ruled out as a source. D. S., A. Preuas, and P. H. Hofschneider, J. Mol. BWZ., 21, 485 (1966). 81 Goulian, M., Z. J. Luw, and A. Kornberg, submitted for publication. *2 Studier, F. W., J. Mol. BWl., 11, 373 (1965). Wang, J. C., personal communication.