Molecular probes are essential to many biological applications including interrogation of tumorigenic polymorphisms, as well as for identifying pathogens of interest. For such large-scale, multiplexed assays where thousands of probes are used, cost can become a limiting factor. Typically, strands in the range of 100—120 bp are made chemically via DNA synthesis automation; however, the longer the strand synthesized, the more failure sequences are produced. Consequently, samples often require PAGE purification to isolate the FLP, which can be very costly. One avenue taken to circumvent this drawback is by enzymatically ligating two shorter constituents of ~equal length to generate the full-length probe. Current methods are generally time-consumptive (~5 hr) and require expensive laboratory equipment and extraction kits to purify the end-product. As such, the focus of our paper has been to detail an alternative method of assembling and purifying ODNs>100 bp, a process that takes just under 15 min (outline given in ).
Figure shows the general process of probe ligation followed by FLP size-exclusion purification.
shows a chromatogram of a crude ligation reaction with A) unincorporated variable and common strands and B) 101 bp FLP (S1–S6) between 5 and 5.5 min. Superimposed on to this image is C) the purified product after filtration through a size-exclusion P-10 polymeric resin, with a final yield three-quarters that of the initial reaction compared to D) standard gel-extraction method, with a final yield at about one-half. Though there are some residual common and variable strands following the size-exclusion purification, it is significantly reduced.
Chromatograms comparing purity and yield of final product before and after purification.
We tested several ligation pools of random variables with different common strands and bridges, and found the overall yields between crude and purified samples comparable to those shown in . However, yield for samples after gel-extraction varied considerably where FLP return was never greater than 50%, and often as low as 25% of the initial full-length amount.
For this work, the process of phosphorylating the common strands was done enzymatically, which required the use of a thermal cycler; however, whether the strands are purchased from an outside vendor or made in-house, the 5′ phosphate group can be coupled directly onto the ODN using a phosphorylation reagent (cheapest of all 5′ ODN modifiers).
While both Fast-Link and Quick ligase mixes were examined for this research, and shown to perform equally well, Quick ligase does have one drawback. When separating out FLP from the unincorporated species through a size-exclusion column, the polyethylene glycol (PEG) in the buffer had a leaching effect on the resin; when hydrated, the sample was opaque, and when lyophilized, a significant amount of precipitation occurred where the remaining pellet could not be dissolved. Alternatively, we did not observe any precipitation with Fast-Link, the buffer of which was absent of PEG. Therefore, this made the latter an obvious choice for our ligation method using size-exclusion purification.
As for modifying the ligation reactions to maximize incorporation of common and variable strands into FLP, we found no benefits of 1) increasing ligation times above 10 min, 2) varying reaction temperatures 3) cycling parameters, 4) annealing samples prior to ligation (e.g. boil, then slowly cool to RT over 15 min), or 5) varying bridge lengths above or below 20 bp; consequently, the one determining factor of ligation efficiency has been to achieve a 1
1 ratio of reactants. And while initial normalization of reactants is essential, it is the most challenging step of the assay; no matter how careful we were in measuring sample concentrations, there was usually some residual variable, bridge and common strand leftover as evidenced by HPLC. This discrepancy can likely be a product of an improperly calibrated pipettor or manual handling error, both of which may be resolved by sample automation. Nevertheless, this has necessitated our use of a method for purifying the FLP from the unligated species as to prevent any possible contamination during downstream application.
Though we determined operational centrifugation time and speed for purifying 101 mers using the P-10 resin, the overall procedure could benefit from further optimization. This includes adjusting 1) resin loading (column resin-bed height and more uniform packaging), 2) RPM, 3) spin time, and 4) resuspending the P-10 slurry in a storage buffer that is [salt] and pH specific. Moreover, cost and time may be further decreased by scaling-up the ligation reactions from 6 variables per pool to 12 or more. The only caution we have is during the purification process; overloading the column with sample could have a negative effect on final yield. To compensate for this, we suggest lowering the final reactant concentrations accordingly.
Here we introduce a quick and cost-effective method for assembling and purifying synthetic DNA probes >100 bp. As an alternative to currently used protocols, our ligation reactions are carried out at RT for only 10 min, and the process of purifying FLP from unligated species is done simply by passing the reaction mixture through a size-exclusion column, which takes less than five minutes. As such, we eliminate the need for costly laboratory equipment including a thermal cycler and gel electrophoresis apparatus; expensive sample extraction kits are also unnecessary. Furthermore, the end-user only requires recycled desalting columns or barrier-style pipette tips for loading with P-10 resin for purifying the ligation reaction. Moreover, for large-scale assays, we envision this method to be fully automated with the use of robotics such as the Biomek FX; here potentially thousands of samples could be pooled, ligated and purified in either a 96, 384 or 1536-well platform in just minutes.