3.3 No self-ligation products from YNNY and RNNR adapters
To prevent self-ligation NNNN adapters, adapters YNNY and RNNR were designed. Adapters ending with YNNY and RNNR can theoretically avoided self-ligation occurred with NNNN adapters. The methylation site dependent adapters have four nucleotides at 5’ extrusion. Their self-ligation between the four wobble NNNN are complementarily paired with 256 types of combinations. When YNNY and RNNR are designed, the Y( C or T base) and Y is not complementary and neither the R and R.
Figure 6 illustrated the results for the three adapters ending with YNNY, RNNR, and NNN with and without T4 ligase. Self-ligated products from NNNN adapters were observed (lane 6). No self-ligation products were observed from the YNNY and RNNR adapters as expected.
Discussion:
The present study developed and tested three strategies to prevent the self-ligation of adapters used in enzyme-mediated methylation sequencing technology. The self-ligation of the two types of methylation site independent adapters have been largely prevented, and more required ligation products with inserts were yielded. The self-ligation of methylation site dependent adapters were completed avoided based on their design.
Adapters targeting the ends yielded from either methylation-sensitive or dependent restriction endonuclease digestion were able to be self-ligated because of their sticking termini designed for ligation with related ends from restriction digestion. These two types of restriction enzymes, particularly the methylation-dependent endonuclease is essential in methylation-dependent endonuclease mediated methylation sequencing. Strategies to prevent adapters’ self-ligation is thus crucial in the development of this new methylation sequencing method (Chinese patent application number 202210881697.2). This new methylation sequencing method may employ four types of adapters, one methylation site-dependent and three methylation site-independent adapters. The self-ligation of methylation site-dependent adapters are completely prevented by eliminating any possible complementary base paring within the four wobble nucleotides in the adapters tested. For the three types of methylation site independent adapters, adapter with 3’ extruded T has no self-ligation issue, but its efficiency in insert coverage of less than 30% [12, 13] prompted us to develop the other two types of adapters tested in the present study. As illustrated in Fig. 1 and Fig. 2, adapters with CGat at its 5’ end together with the enzyme ClaI virtually changed an reversible reaction to unidirectional reaction of insert-adapter ligation. For the adapter with terminal TAat string, the combination of enzymes MseI and AseI substantially decreased adapters’ self-ligation (Fig. 3).
As illustrated in Fig. 1, an ideal strategy should have three effects: i. to inhibit adapters’ self-ligation, ii. to promote the production of insert-adapter ligation, and iii. to regenerate adapters from adapters’ self-ligation products. The present study demonstrated three types of strategies in decreasing or preventing adapters’ self-ligation, with different advantages and disadvantages. The strategy in preventing CGat adapters’ self-ligation is overall better than the other two as it is fulfill all the three standards as discussed.
The insert-adapter ligated products and adapters’ self-ligated product in methylation dependent endonuclease mediated methylation sequencing technology are hardly separated based on their sizes as they may be predominantly between 100 and 200 base pairs[14-16]. In real methylation sequencing library, the size difference between a large portion of ligation products of insert and adapters and adapters’ self-ligation is about 30-34 bp, and some are as less as 17 bp. For example, when a palindromic methylation site is cut by MspJI ((N)13mCGmCG(N)9), the insert in the library is only 30 bp. Application of more types of adapters to decrease or eliminate self-ligated adapters has at least four advantages. Firstly, the three types of methylation site-independent adapters can be combined to increase the coverage of inserts to be assayed as more inserts are ligated with adapters. Secondly, insert-adapter ligation is preferred over insert-insert ligation and thus coverage of insert is further increased. Thirdly, less self-ligated adapters in the final sequencing library will minimize sequencing noise. Fourthly, decrease self-ligated adapters will greatly cut the cost of methylation sequencing by maximally using the sequencing chip with most of the clusters are formed from insert-adapter products. These advantages are precious in methylation sequencing of single cell and circulating free DNA samples.
In conclusion, three strategies were introduced in our newly developed methylation-dependent endonuclease mediated methylation sequencing method. Self-ligation of the methylation site-dependent adapters could be theoretically eliminated by design. For the two methylation site-independent adapters, the one with methylation sensitive enzymes is virtually abolished by forcing the reversible reaction to a unidirectional reaction of insert-adapter ligation. The third type of adapter with methylation unrelated endonuclease, the combination of enzymes MseI and AseI as an example showed its usefulness in decreasing adapters’ self-ligation. Although these three strategies are crucial for the development of methylation-dependent endonuclease mediated methylation sequencing, they are not equally efficient and improvement is expected whenever new suitable endonucleases are available. These strategies may find more applications in the development of other enzyme-mediated genomic sequencing technologies.