NUCLEIC ACID-PROTEIN RECOGNITION New York: Academic H.J. Vogel (Ed.) Press, 1977, 261-268 -'T4 Ligase Joins Flush-Einded DNA Duplexes Generated by Restriction Fndonucleases S. D. EHRLICH,*? V. SGARAMELLA,t AND JOSHUA LEDERBERG* *Depariment of Genetics Stanford University Medical Center Stanford, California + Laboratorio di Genetica Biochimica ed Exoluzionistica del CNR. Pavia, Italy The resealing of single-strand interruptions (nicks) in double- stranded DNA molecules is catalyzed by appropriate polynucleotide ligases (1). Prokaryotic and eukaryotic cells have been shown to con- tain such enzymes: so far the most extensively studied are the en- zymes purified from uninfected and from T4-infected Escherichia coli cells. The former have received much attention, mainly thanks to the work of LB. Lehman and co-workers (1). But T4 ligase displays addi- tional properties of great interest, such as the ability to join DNA-RNA hybrids (2) and flush-ended DNA duplexes (3). The interaction between the nicked DNA and the enzyme takes place after the enzyme has been activated in the fonn of an enzyme-adenylate intermediate: the activation is not dependent on DNA. The activated complex recognizes a nick between a 5’- phosphoryl! and a 3'-hydroxyl group (4). Both of these functions are necessary for the ligation (1) and have to be kept in close register bya complementary continuous strand. The activated enzyme then links ’ Present address: Institut de Biologie Moléculaire, Faculté de Science, Paris, France. 261 oe ee EE AE OT 8 Seat PEE RRS Bre mene Ame Ly ieeibaela anal: 2 Se ae eter nn nes _ i: | | | | is 262. s. D. EHRLICH ct al. : r ; ia its adeny! moiety to the 5'-phosphate through a pyrophosphate bona. 12 ie The phosphodiester bond linking the two adjacent mucleotides 1s iz t eventually formed and is accompanied by the stoichiometric release ¢ of AMP. All of these events have been established for both the E. coli and the T4 ligase and are thoroughly reviewed by Lehman (1). The recognition of DNA by the ligase thus takes place between an activated enzyme and anicked double helix: the discovery that the E. coli ligase in the presence of AMP can catalyze the reverse reaction and convert supercoiled DNA into relaxed and nicked circles (5) is indicative of the affinity this ligase also has for continuous supercoile DNA double helixes. Data on the reversal of the joining reaction by _the T4 ligase are regrettably missing, but its affinity for nicked double-helical DNA molecules is nevertheless well documented (6). How can the T4 ligase join head-to-head flush-ended DNA du- plexes? A brief review of the information available on this reaction is certainly useful in view of its theoretical and practical interest. The so-called “terminal” joining ability of the T4 ligase was discov- ered using segments of the synthetic alanine transfer DNA gene pre- pared and characterized in Khorana’s laboratory (3); the nearest- neighbor analysis of the joined products gave unequivocal evidence that the ligation had occurred at the fully base-paired tennini of two opposed duplexes (7). The limited availability of such substrate al- lowed only to reach the conclusion that it takes place ata reasonable rate and with a yield comparable to those observed in the “cohesive” joining of short duplexes held together by single-stranded ends 4-6 nucleotides long (3,7). Among the naturally occurring substrates with putative base-paired ends, the DNA of Salmonella typhimurium bacteriophage P22 was se- lected for practical reasons. Its native DNA cannot be terminally an- nealed except after critical portions of the duplex, corresponding to the terminal repetitions, have been converted into single strands by means of lambda exonuclease or E. coli exonuclease III (8). Intact P22 INA could be joined by the TA ligase, although not at very high effi- ciency: in a rather slow reaction, the T4 ligase converted 30-40% of the DNA molecules into linear dimers, trimers, and higher oligomers, as based on sucrose gradients and electron microscopy analysis (9). The E. coli ligase was unable to perform the same reaction on p22 DNA, but an exciting by-product of this investigation wes the discov- ery of the cohesive nature of the termini produced by the restriction endonuclease EcoRI! (9). The explosion of the research on other restriction enzymes has now made available a rich supply of substrate {or the terminal joining reaction. Among the various flush-ended du- 2 mene yt qh men mt ised tei ote sets a0 og lng epaatane® ‘ony care EGET RS I EERE SES REIN OE rN Le ee ee ne = O by 6). ju- 1is ov- are- ast- nce two al- ible ve” 4-6 ired 3 Se- ran- ig to s by p22 effi- % of ners, s (9). p22 $cOv- ction other strate d du- Beni gh iin el Shep ee tee bo 4 LIGASE JOINS FLUSH-ENDED DNA DUPLENES 265 with positive results, the most useful turned out to be plexes analyzed those produced by the Bacillus subtilis R endo (10). In the DNA of the SPP1 phage, this enzyme introduces about 100 cuts (10,11), and we have used these flush-ended molecules to develop a new assay for the terminal joining reaction. This assay is based on the intramolecular circularization of the segments in the presence of the T4 ligase and on the visualization of the circles with the electron microscope. Figure I gives an example of what one sees after aqueous spreading ofa ligated sample of BsuR endo-generated segments (the details of this assay will be presented elsewhere). , Z w = = Bs Sees Fig. 1. Electron microscopic visualization of the circles formed by the T4 ligase on flush-ended segments produced by BsuR endo on SPP1 DNA. Purification of BsuR endo and its use were essentially according to Bron et al. (10). For the purification of th T4 ligase, the protocol of Weiss ct al. was followed until fraction V (13). e tity os gtx baa 264 Ss. D. EHRLICH et al. are the efficiency of joining DNA Gu through flush-ended termini to that taking place through the termini introduced by EcoRL into SPP} DNA (10). Ti is important that the average size of the linear molecules produced by limit digestion with EcoRI is about 3 times longer than that of the limit digest products of BsuR on the same substrate (10). We therefore resorted to the use of limited digestion of SPP1 DNA with BsuRendo, It seemed interesting to comp plexes short cohesive so that the segments could be of similar average size as those gen- erated by EcoRI on SPP1 DNA. In Fig. 2 are given the kinetics of circul seginents: at all the temperatures tested; r at comparable rates. The fi arization of E coRI-generated anging from 5° to 42°C, the reaction proceeds nal extents are also very 70 60 50 40 30 PERCENT CIRCLES 20F 10 en ee ee ee ee ° 0. 30. «+40 «50% “ 90 0 10 TIME (MIN) ature on the T4 ligase-catalyzed joining of cohesive- oRI on SPP1 DNA. A reaction mixture of 720 pl con- Fig. 2. Effect of the temper ended segments generated by Ec taining 1 wg of EcoRl cut DNA per millititer, 50 m\f Tris-Cl, pH 7.6, lmif 2-mnercaptoethanol, 50 pM ATP, 10 pM nucleotide tRN V T4 ligase was prepared at 0°C and divided into five portion at the various temperatures. Aliquots of 20 pl were withdrawn transferred to chilled tubes containing 1 pl of 0.5 M EDTA, and he min. For electron microscope analysis, the samples were spre and Schnés (12) and observed in a Philips 200. For e were scored to determine the percentage of circles. s, which were incubated 5mM Mg Ch, A, and 0.6 unit of fraction at the indicated times, ated at 65°C for 5 ad according to Inman ach point, at least 100 molecules ce nee em RR ER LI eek NAR cae A AEM MO 2 ARTIS TEN eS RL TL IT EEO ER EAP TE RE BS poe” | { . ~ T4 LIGASE JOINS FLUSH-ENDED DNA DUPLEXES 265 Jf 60 1 I i T TF fT 50F- 25°77] n _y/ ve) B Wo 40h 4 O° e— —@ 159 « —— TF s oa — Bb 5 Pr , “ wer fa 35°) wi / _*. © 20 foe = . 4 a n/ oe TH a SF 0 1? , | 0 mo a , 10 7 4 7. o J {I L/Z ] I 0 70 20 30 40 50°% 210 TIME (MIN) Fig. 3. Effect of the temperature on the T4 ligase joining of flush-ended DNA seg- ments. A reaction mixture was set up as described in Fig. 2, except that it contained per milliliter 1 xg of Aush-ended DNA segments generated by BsuR enda on SPP1 DNA and 2 units of ligase. close, the highest level being probably attained between 15° and 25°C. Figure 3 shows the results of an analogous experiment on Bsu-generated segments. It is apparent that here the highest rate and extent are obtained at 25°C, and the lowest at 42°C. On the basis of these results, we then investigated the effects of various amounts of enzymes on both types of joining. The enzyme activity was deter- mined by the ATP—”PP, exchange reaction (13) as shown in the inset to Fig. 4. Figure 4 gives the results obtained with EcoR\-generated segments. The initial rates seem to be slightly affected by different amounts of enzyme, whereas the final extents tend to level at different values, in spite of prolonged incubations. In the reaction with flush- -. ended substrates, the differences are even more pronounced (Fig. 5). Both initial rates and final extents are approximately proportional to the amount of enzyme present. As compared to the cohesive joining, te n the initial rates are close to fiftyfold lower. This difference can be ex- d plained in several ways. One explanation is that the presence of cohe- _ sive single-stranded termini allows the interaction between two du- plexes to last long enough for the enzyme, even when present at rela- tively low concentrations, to bring forth a rapid ligation. Flush-ended termini, on the other hand, could interact by means of stacking forces, : rae FESSOR ENS RT CPE Aan a PO BE LE SN Te re TI ORE IE ara RT en Ghd neat OMe eT 266 S$. D. EHRLICH et al. “ PERCENT CIRCLES 3 al LIGASE I ! 1 1d i_/ /I 10 20. «30~+«40 50 60? 150 TIME (MIN) Fig. 4. Effect of the amount of enzyme on the joining of EcoRI-generated segments. A500-pl mixture was set up as in Fig. 2 except for the absence of the enzyme. The mix- ture was divided into five 100-pl aliquots, and then appropriate dilutions of the enzyme were added corresponding to the units indicated next to the curves. The inset gives the linear response of the T4 ligase in the ATP-“PP, exchange reaction (13). The joining reactions were run at 20°C. but the apposition would conceivably be very transient. Tempera- tures higher than the optimal one could destabilize the stacking in- teraction, and lower ones could affect the frequency of productive col- lisions. The multistage nature of the joining reaction has already been dis- cussed: in the presence of flush-ended termini, some of the intermedi- ate steps could be slowed down. For example, in the ligation of the small duplexes intermediate in the assembly of the synthetic alanine-tRNA gene (14), small changes in the parameters of the reac- tion (temperature, Mg** or ATP concentrations) have been found to- cause drastic differences in the extent of ligation: in two cases (14,15) of cohesive joining it has been possible to isolate the adenylated oligonucleotides responsible for the low level of ligation. In both cases the interruption to be sealed was six base pairs away from the flush terminus of the duplex under investigation. In the other cases, where different plateaus were obtained, the reasons have not been found, except for a possible not better defined “freezing” of the sub- SL 60 I T a T fo O 0.4 UNITS 50 6 0.2 UNITS _ OQ 0.1 UNITS LO n A 0.05 UNITS we 40 gE 0.025 UNITS 4 oO « oO , 30h 4 ca & 20 ? /, ao 8 27? wh JA O- ee go WX a ; { ! ! L/ TIME (MIN) Fig. 5. Effect of the amount of T4 ligase on the joining of Bsul-generated flush- ended DNA segments. The mixture was set up as described for Fig. 4 except for the presence of flush-ended DNA segments. strate in some unreactive structure. Experiments are in progress aimed at establishing whether the substantial amount of uncircu- larized substrate left after incubation at suboptimal temperatures or in the presence of lower levels of enzyme has been converted into the - adenylated form. An additional explanation for the low efficiency of : the terminal ligation is the possibility that the enzyme is inactivated = faster in the presence of flush-ended substrates than of cohesive ter- | mini. fe In order to shed light on these possible explanations, experiments i fh Cee ka teeny ae i- are planned in which the efficiency of joining flush-ended duplexes is e compared to that in which the cohesion is mediated by different a c single-stranded termini of decreasing length. » ACKNOWLEDGMENTS £ -h We want to thank Drs. K. Murray and S. Bron for initial gifts of BsuR endo, and the i e same researchers and Dr. H. G. Zachau for making available information on the purif- cation of this enzyme before publication. This work was made possible by grants from the National Institute of Health to Dr. J. Lederberg and from the U.S.A.-Italian National Council of Research (C.N.R.) Coopera- a five Program. De pe ne eee a . a . : os cy ERTS: arene: a gee sa rie Raped “pyeerer Sbbpone fig Saag LAST eS hae BSS eee amet prea} 26 DOnNnAans 14. 15. 8 S. D. EHRLICH et al. REFERENCES . Lehman, I. R. (1974) Science 186, 790. _ Fareed, G. C., Wilt, E. M., and Richardsen, C. C. (1971) J. Biol. Chem. 246, 925; Kleppe, K., van de Sande, H. J., and Khorana, H. G. (1970) Proc. Natl. Acad. Sci. U.S. A. 67, 68. . Sgaramella, V., van de Sande, H. J., and Khorana, H. G. (1970) Proc. Natl. Acad. Sci. U.S. A. 67, 1468. . Gumport, R. L, and Lehman, I. R. (1971) Proc. Natl. Acad. Sci. U.S. A. 68, 2559 . Modrich, P., Lehman, 1. R., and Wang, J. C. (1972) J. Biol. Chem. 247, 6370. . Weiss, B., and Richardson, C. C. (1967) Proe. Nail. Acad. Sci. U.S. A. 57, 1021. . Sgaramella, V., and Khorana, H. G. (1972) J. Mol. Biol. 72, 475. . Sgaramella, V., Ehrlich, S. D., and Lederberg, J. (1976) J. Mol. Biol. 105, 587. . Sgaramella, V. (1972) Proc. Natl. Acad. Sci. U. S. A. 69, 3389. 10. Ly. 12. 13. Bron, S., Murray, K., and Trautner, T. A. (1975) Mol. Gen. Genet. 143, 13. Sgaramella, V. (1976) Atti Assoc. Genet. Ital. 21, 16. Inman, R. B., and Schnés, M. (1970) J. Mol. Biol. 49, 93. Weiss, B., Jacquemin-Sablon, A., Live, T. R., Fareed, G. C., and Richardson, C. C. (1968) J. Biol. Chem. 243, 4543. Svaramella, V., and Khorana, H. G. (1972) J. Mol. Biol. 72, 427. Van de Sande, H. J., Caruthers, M. H., Sgaramella, V., Yamada, T., and Khorana, H. G. (1972) J. Mol. Biol. 72, 457.