Total Synthesis of the Structural Gene for the. Precursor of a Tyrosine Suppressor Transfer RNA from Escherichia coli OF SEGMENTS

The polynucleotide ligase-catalyzed joining of the eight chemically synthesized deoxypolynucleotides (segments 19 to 26), comprising the nucleotide sequence 86-126 of the DNA corresponding to the Escherichia coli tyrosine tRNA precursor has been investigated. Joining was studied using various combinations of 3, 4, or larger number of segments at a time. The the three-component the Joining of the five-and six-component systems satisfactory with yields to 60%. The three duplexes [IVa] in step reactions in yields were characterized. the failure rest

papers. Unfortunately, a contamination was discovered in the 5'.end group of segment 20. When the latter was phosphorylated with [y-32P]ATP and polynucleotide kinase and the product analyzed for radioactive 5'.nucleotides, about 10% of radioactivity was found in pdA, the remainder being in pdG as expected. The implications of this in the experimental work are described below. Methods-These were as described in an accompanying paper (8).

Joining Experiments
Using Three Deoxyribopolynucleotide and purification of the joined product using segments 20, 21, 22, and 23. Segments 21 and 22 were phosphorylated using 32P of the same specific, activity. Two reaction mixtures (22 pl each) were set up containing Tr.is-HCl (pH 7.6) (20 m@, dithiothreitol (10 mM), ATP (66 jtM), and each of the four segments of 1 PM concentration. The concentration of MgCl, was 10 and 20 mM respectively and ligase was added at 1000 units/ml. When the reactions (0') had leveled off (at 30 to 35% joining), excess EDTA was added and the pooled mixtures were fractionated at 4" on a Bio-Gel A-0.5m column (105 x 1.2 cm) using 0.1 M triethylammonium bicarbonate at 4O. Fractions of 300 pl were collected and analyzed for 92P radioactivity. The numbers in parentheses after the counts per min are the experimental molar ratios. consisting of segments 19 to 23 gave an optimal yield of 39% at 20 mM Mg2+ ion concentration. At lower Mg2+ ion concentrations (5 or 10 mM) the yield was reduced to 25%. System 9 (segments 20 to 24) gave a much higher yield (53%), the concentration of Mg*+ ions being 10 mM. At 20 mM Mg*+ ion concentration, there seemed to be a further increase (to 57%) in yield, but there was a sharp decrease (to 38%) at 5 mM concentration.
The five-component system (system 9), consisting of segments 20 to 24 was studied further by isolation and characterization of the joined product. The reaction was carried out with segments 21, 22, and 24 carrying 5'-S*P-phosphate groups and segments 20 and 23 being unphosphorylated at the 5'-ends. The elution pattern is shown in Fig. 4. Only one main peak which corresponded to the joined product was found, as well as the peak corresponding to the unreacted oligonucleotides.
The characterization is in Table III. The phosphatase assay indicated 96% of the 3aP was resistant to the enzyme, and, therefore, present in internal positions as expected. The 5'-analysis verified that the main product corresponded to the duplex expected from the five components. Equivalent counts were found in pdA, pdG, and pdT. Tris-HCl (pH 7.6), 5 mM dithiothreitol, 75 FM ATP, and 1 PM each segment. The reaction mixture was incubated at 37" for 1 hour, cooled to O", and 5 units of ligase added. The reaction after 16 hours showed that 57% of the total radioactivity had become resistant to the phosphatase. The reaction was terminated with an excess of EDTA, applied to a Bio-Gel A-0.5m column (105 x 1.2 cm) and the column was eluted with 0.1 M triethylammonium bicarbonate at 4O. Fractions of 300 ~1 were collected and analyzed for radioactivity. The numbers in parentheses after the counts per min are experimental molar ratios. 19 to 24 was obtained in 42% yield at 10 mM magnesium ion and at 0" ( Table I). The yield could be improved slightly (48 to 50%) when the magnesium chloride concentration was increased to 20 mM. At 5 mM magnesium ion, the yield was only 30%. These results were not affected by the method of annealing. Segments at the appropriate magnesium concentration were either heated to 100" or taken to 37" followed by slow cooling at 0' without a significant difference in the yield of the duplex. The products of reactions carried out at 10 and 20 mM magnesium ion were isolated from small scale reaction mixtures. The elution patterns were similar to that shown below in Fig. 5 for the large scale reaction. Analyses of the isolated duplexes were all in agreement with the expected values. Therefore, the duplex [IVa] could be prepared in a reasonably satisfactory yield in a one-step joining reaction containing the six segments 19 to 24. in Peak I is given in Table  IV. Degradation to 3'-mononucleotides gave radioactivity in dTp and dCp in a molar ratio of 1:l. Analysis by digestion to 5'-mononucleotides gave radioactivity in pdA, pdG, and pdT, the molar ratio being 2:1:1, respectively. All these results are as expected for the correct duplex. Further, the duplex (Peak Z, Fig. 5) was analyzed on polyacrylamide gels under denaturing conditions (Fig. 6). At 2.8 pM concentration, the duplex [IV] gave a diffuse pattern as if denaturation was partial. However, after phosphorylation of the external 5'-hydroxyl groups with [3*P]ATP of high specific activity, it was necessary to use only a low concentration (20 nM). Under these conditions, the duplex was completely denatured and the separated strands were clearly visible. No faster traveling, partially joined bands were visible after phosphorylation under denaturing conditions. Polyacrylamide gel electrophoresis also indicated that Fractions II and III (Fig. 5 Fig. 7 and the conditions of the experiment are shown in the legend. The joined product was separated satisfactorily as shown in the figure and the pooled peak was characterized as in Table V. All of the results of nearest neighbor analysis and the Y-nucleotide analysis were as expected for the specific joinings. Thus, the phosphatase assay showed the joined product to be 96% resistant. On degradation to 3'-nucleotides, radioactivity was found mostly in dGp, dTp, and dCp in the ratio of 1.07:2:1.9, the expected ratio being 1:2:2. Analysis of 5'-nucleotides gave radioactivity in pdA, pdG, and pdT and the molar ratio was 2:2.26:0.83, the expected ratio being 2:2:1. . Both reaction mixtures were incubated at 37" for 10 min, then cooled slowly (about 1 hour) to 4'. ATP (40 FM), dithiothreitol (2 mM), and ligase (600 units/ml) were added and the mixtures kept at 4". After 15 hours, phosphatase assay showed that the extent of radioactivity resistant to the phosphatase was 37.7% and 38.8%, respectively, in the two reaction mixtures. Electrophoresis of the products on a 15% polyacrylamide gel showed identical mobility. Only segment 25 was concluded to have joined and the extent of this joining seemed to correspond to the extent of phosphorylation at the 5'-ends in the above preparation of duplex [IVa].
Preparation of Duplex [IVc] (Segments 20 to 25) with Unphosphorylated 5'-OH Groups-Phosphorylated segments 21 to 24 (270 to 300 pmol of each), nonphosphorylated segments 20 (350 pmol) and 25 (465 pmol) in 50 mM Tris (pH 8.0) and 20 mM MgCl, were heated for 5 min at 80" and then cooled to 4". ATP (0.1 mM), dithiothreitol (2 mM), and ligase (130 units/ ml) were then added, the final reaction volume being 0.125 ml. The reaction, which was carried out at 4', was followed by the alkaline phosphatase assay. The resistance of 3zP radioactivity increased from 4.5% at zero time to 51.2% at 1.5 hours, 55% at 7 hours, and 58% at 26 hours. Thus, the kinetics were similar to those described below in the experiment of Fig. 10. The total reaction mixture was applied to a 15% polyacrylamide gel and the pattern obtained is shown in Fig. 8. The band corresponding to the duplex [IVc] was eluted from the gel with 2 M triethylammonium bicarbonate. The product had mobility identical on the 15% gel with that prepared below using [32P]segment 20. Further, phosphorylation with [y-33P]ATP as described below proceeded as expected.

Phosphorylation of Duplex [IVc] (Segments 20 to 25) Containing 5'-OH Groups with [Y-~~P~ATP and Polynucleotide
Kinase-The reaction mixture in 18 ~1 contained 30 pmol of the duplex prepared above, 130 pmol of [Y-~~P]ATP, and the standard components for the kinase reaction including 2 mM spermine. After addition of the kinase (final concentration about 50 units/ml), the reaction was followed by the DEAEcellulose paper assay. The per cent of total radioactivity remaining at origin at different time intervals was: 10 min, 29%; 20 min, 37%; 40 min, 38.6%; and 60 min, 39.8%. Theoretical expectation for utilization of asP from ATP was 46%. The product after isolation was analyzed by degradation to 5'mononucleotides.
All of the 33P radioactivity was found in pdG.  Preliminary experiments using equimolar amounts of the five phosphorylated segments ( [3SP]segment 20, 32P-phosphorylated segments 21 to 24) and of unphosphorylated segment 25 showed that, (a) the joining of segments 20 to 24 to form the duplex was rapid but that the joining of segment 25 to segment 23 was slower; (5) some of the Y'-end group in segment 20 was converted to a form resistant to the bacterial alkaline phosphatase; and (c) electrophoresis on polyacrylamide gels showed the formation of a higher molecular weight product, which appeared to be the dimer (Fig. 9). (The characterization of this dimer is given below in Fig. 10 and Table VI.) The formation of the dimer could be inhibited by including an excess of nonphosphorylated segment 19, which itself was unable to undergo covalent joining. Further, the joining of segment 25 to segment 23 in the duplex could be accelerated by using an excess of segment 25 relative to the other segments. Following these observations, the synthesis of duplex [IVc] containing segments 20 to 25 ( [33P]segment 20) was performed as shown in Fig. 10.
5'-92P-phosphorylated segments 21 to 24 (4000 pmol each), [SSP]segment 20 (4100 pmol), and 2-to 2.5-fold excess of each of segments 25 and 19 were used in a joining reaction as described in the legend to Fig. 10. The kinetics of the joining are shown in the inset. The products were separated by gel filtration as shown. Before pooling the peaks, especially of the desired material, small amounts of selected fractions were examined by polyacrylamide gel electrophoresis (Fig. 11). This analysis helped delineate the dimer, the required duplex containing all the six segments and the duplex lacking segment 25.
Characterization of Peak A-This was concluded to be the dimer with the structure shown in Fig. 9. Thus, its molecular weight, as estimated from electrophoretic mobility on a polyacrylamide gel was consistent with the structure. All of the analysis (Table VI) including 32P/33P ratio (4.11) and the 3'-nucleotide analysis (80% of s3P in dCp) were all in agreement with this conclusion. Further, it has been mentioned above that the 5'-end nucleotide [S3P]guanosine) in segment 20 was contaminated by [""P ]pdA in the amount of 8%. When the dimer was analyzed for 33P radioactivity after degradation to 5'-nucleotides, a high proportion (26%) of 33P radioactivity was found in pdA. This enrichment of segment 20 containing pdA at Y-end in the dimer was to be expected because a standard A:T base-pair would now be formed instead of a G:T base-pair (Fig. 9). Finally, it is interesting to note that gel electrophoresis performed after heating and chilling of the dimer showed a single product with mobility similar to that of duplex [IVc] (Peak B of Fig. 10). This result is interpreted to mean that the strands of the dimer after coming apart simply form hairpins on themselves and the product, therefore, having one-half the molecular weight has the mobility close to that of duplex [IVc 1.
Characterization of Peak B-This contained the required duplex [IVc]. Thus, its electrophoretic mobility was consistent with its size (Fig. 11) and all of the analyses (Table VII) were in agreement with those expected. The molar ratio of a2P/aaP (terminal label on segment 20) was found to be 3.97, theoretical being 4. On degradation to 5'-nucleotides, a2P was found in pdA, pdG, and pdT, the molar ratios being 1.0:2.08: 0.92; those expected were 1:2:1. 39P was found mostly in pdG but about 4% was also present in pdA. This reduced "contamination" by pdA was evidently due to the preferential participation of the duplex [IVc] carrying 5'-pdA end group in dimer formation. On degradation of [IVc] to 3'-nucleotides, dCp, dGp, and dTp contained ?P in the ratio of 1.0:0.99:2.18, the theoretical ratio being 1:1:2. These analyses taken together   showed that the addition of the latter segment inhibited the formation of the dimer (Fig. 9)    of the duplex [IVa], an analysis of the different parts of the product peak (Fig. 5) was similarly made to separate contaminants lacking terminal segments.
The use of prephosphorylated segment 20 in the synthesis of the duplex [IVc] gave interesting results. First, the formation of the dimer (Fig. 9) confirmed what was already concluded in the preceding paper (1) that the ligase can with facility bring "head to head" dimer formation even if the base-pairing of the protruding tetranucleotide sequence is not perfect. Second, it was interesting and useful to be able to inhibit the dimer formation by "tying up" the singlestranded region with segment 19 containing the complementary sequence. Although prephosphorylated segment 20 could be used reasonably well in the synthesis of duplex PVC], the use of prephosphorylated segments at termini cannot be recommended.
Instead, the aim should be to so arrange the joining reactions that the 5'-OH groups, which would be available for subsequent phosphorylation, are carried by those terminal segments which have protruding singlestranded ends.
While, as discussed above, the segments at termini may have a slower rate of joining, the experience with the joining of segment 26 was a surprise. Indeed, the results in these experiments emphasized again the lack of basic understanding in this area of the ligase-catalyzed joining of short deoxyoligonucleotide segments. The hexanucleotide segment (segment 26, Fig. 1) failed to join to form the duplex [IVd] in all the many attempts that were carried out. This failure was surprising in view of the past successful joining of hexanucleotides at termini. At least two successful examples are: (a) the joining of p-C-C-A-C-C-A (segment 2) to the remainder of duplex [A] in the synthesis of the gene for yeast alanine tRNA (ll), and (bj the joining of segment 1 (3'-G-G-T-G-G-T-5') to segment 3 in duplex p]' described in a preceding paper (8). Further, it is striking that in the alternative ap-proach, in which prior joining of segment 24 to segment 26 was attempted, careful annealing of the three components (segments 24 to 26) was essential. This is one of the few joining reactions studied so far in which the necessity for slow annealing has been demonstrated.