The bands of C3p and C5p are abnormally pale because their sequences with high ratio of pyrimidine are hard to be stained.
As we can see, the bands in lane 7 and 8 are darker than those in lane 3 and 4. There is more C3p left than C5p, suggesting that the productivity of AC5 is higher than that of AC3. We suspect that as the length of T4 DNA ligase binding domain is 11bp and the linking domain of our catenanes are only 10bp, the TTT hinge is too short. It may cause steric hindrance for the combination of T4 DNA ligase, thus reduce the productivity.
(The samples of two electrophoreses are from the same reaction system.)
The mobility of catenane depends on its topology. Theoretically, the bigger linking number (LK)is, the higher its mobility is. However, in the gel of different concentration, the order of them changes, so we used two different gels to reveal their linking numbers. According to the former research , we can be sure that in the lane of AC52, AC51 and AC50, the order of linking number in the 10% denaturing PAGE from top to bottom is LK=2\LK=3\LK=1\LK=4\LK=5. But the mobility of the second bands, LK=3 and LK=1 were reversed in comparison with the 14% gel. In our ideal design, CA5+ is supposed to have minimal topological complexity, whose linking number is one. Because CA5+ structure contains two staple stands which can control the topological structure of two interlocked rings. Furthermore, the cyclization of adjacent interlocked DNA rings depends on the oligonucleotides we added which played the role of splints, leading to less probability of improper intertwining compared to the cyclization with the help of existing DNA rings. Apparently, CA5+ appears in the LK=1 position in the gel maps, verifying the previous inference. Above all, we can conclude that more staple strands and oligonucleotides can make great contribution to the formation of desired topological structure, which means less linking number and more right products.
We use three scaffolds to find the best of them, L1\L3\L5. At last, L3 turns out to be superior to the others in productivity. As we can see, the productivity is varied with the increasing length of linking domain and the decreasing length of hinge. According to the gel map, the productivity of L3 is the best. We speculate that compared with L1, the increase of L3 product yield may because the linking domain of L3 is longer, which promotes T4 DNA ligase binding with it. Compared to L3, L5 has lower yield, which is caused by the steric hindrance which brings about more difficulties in the intertwining of two neighboring strands. Hence, the scaffold L3 is more suitable than L1 and L5 for the construction of the four-ring catenanes.
In the gel map, there still are bands of scaffold even after digestion, because the scaffold is attached to four rings and exists as double helix, which cannot be digested by EXO-I. In conclusion, L3 is the best choice, even though the bands in the gel map cannot be distinguished clearly from A5p and B5p.
In this part, we put forward two possible methods to achieve high efficiency of catenane synthesis. The scaffold strands and artificially added oligonucleotide strands are respectively function as the splint to compare their different abilities to cyclize the ssDNA strands. Experimental results show that the former one is a better choice. As we can see, there are more topologically byproducts in AB, while there are more byproducts of different ring number in AsB, ABs and AsBs. In AB, only the strands that bind to the scaffold can finally get cyclized, while unfree linking domain leads to incorrect intertwining. However, with the help of oligonucleotide splint, two-ring and three-ring catenanes can be produced, and even the cyclization of the fifth strand can still be achieved without the combination of scaffold. So we can only observe the five-ring catenanes in the oligonucleotide splint existing experiment.
It suggests that the addition of staples can indeed reduce the by-product whose topology is improper. As previously said, there are five possible Lk between two interlocked catenane, and there will be more topological isomers among four circles. Luckily, the presence of spAs and spBs successfully reduce the main product bands from six to four. As the general rule goes, the smaller LK is, the less mobility of the products is. According to the gel map, the fact that the mobility of the products decreases which is accordant to the rule indicates that the LK of our product decreases. Overall, AB is the best for the construction of four-ring catenane, so we chose AB as the first layer to form the second layer.
After the formation of the first layer, we put our hands to building the second layer. We adopt two methods to fabricate the second layer. In method one, the first layer serves as splints, while in the other artificially added oligonucleotides do. The result demonstrates that the former method is infeasible. The possible reason might be the steric hindrance generated by the interlocked rings in the first layer, which is likely to induce inhibition of the ligase by blocking the combination with linking domain. Only applying the second method, can we observe the formation of aimed catenanes.
In this figure, there are two-ring catenanes appears because we didn’t extract the four-ring catenanes out from the first layer solution system but chose to add excess C5p and D5p in order to improve the recovery and save time. It seems that remnant circle A or B will disturb the formation of the second layer, as the two ends of C5p or D5p may attach to different circles and can’t be cyclized. Although we added the staple strands to facilitate the formation of right eight-ring catenanes, staples still cannot make it. However, staple strands can only promote the formation of AC5 or BD5.
CsDs is a negative control group and set to avoid disturbing of the polycatenane composed of circle C and D.
There is larger catenane productivity in the lane of R2CsDs, but few products in the lane of R2Cs and R2Ds, perhaps because intertwining between C5ps and D5ps can promote the cyclization of each other. As for the difference between C5p and C5ps, the linking domain of C5p is 5bp+5bp. But when it’s comes to the C5ps, the binding domain of C5ps is 10bp and the linking domain is 12bp, which means that C5ps can provide enough space for the ligase combination. So in this complex system, C5ps is better than C5p in productivity.
We demonstrated that staples and splints can indeed prevent the topological byproduct because they can bind to complementary single strands and make them harder to pair or intertwine with others. What’s more, when using scaffold to instruct the formation of catenane, both the length of linking domain and hinge should be taken into consideration, as the former determines the linking number and the binding efficiency of T4 DNA ligase, and the latter determines the chance of adjacent circles to intertwine and the steric hindrance against ligase. Using the bottom of the first layer as splint failed now, but may be achieved if the circle is lager with less steric hindrance. Using an oligonucleotide as splint succeeded, and if we go on, larger catenanes may be produced.
Monitoring Single-Stranded DNA Secondary Structure Formation byDetermining the Topological State of DNA Catenanes.