DNA Separations with Controlled Trans-membrane NaCl Gradient

In an attempt to minimize flow due to salt gradients, we ran an experiment in which an intentional gradient was used to set up a slight counterflow from the filtrate to the retentate.  This was accomplished by having a 55mM NaCl concentration in the retentate and a 45mM NaCl concentration in the filtrate.  The top drop was 9 µL in volume while the bottom drop was 11µL in volume.  This experiment was performed in our Teflon blocks and a 50 mM NaCl reservoir was used to control ambient humidity in the chamber.  Membranes were from wafer 101 and were reused from previous experiments.  We obtained three replicates for this experiment.  A 10 base pair (bp) DNA ladder was used, which begins at 10 bp and steps up to 330 bp in steps of 10 bp.  Filtration time was 2 days.  For each filtration the retentate and filtrate were run on a 4% agarose gel and imaged with Syber Gold nucleic acid stain.  The following graphs are of filtrate/retentate ratios for DNA bands of varying sizes expressed in base pairs (bp).  Above 100 bp, groups of bands were averaged due to poor resolution.  The bp value for bands above 100 bp are the average size of the group of bands.

We believe that #1 had obstructed pores.  Since this is a recycled membrane this is not too surprising.  In our lab, we only have the ability to check if membranes have pinholes.  Experiments #2 and #3 were fit quite nicely by our exponential decay plus plateau.  R2 values for these two data sets were around 0.94. Given this data, I chose to treat experiment #1 as a bust and averaged the results of #2 and #3.  This yielded the following:
As compared to other data sets, our error has been greatly reduced.  It appears possible to reduce the error even further given that much of the difference between #2 and #3 is systematic, with the ratios in #2 being about 0.2 less than for #3.  We hope to generate more replicates with favorable results and constrain our standard deviations further.  To accomplish this we will need another wafer of membranes with similar characteristics to #101.  You will notice that our cutoff at ~50mM NaCl concentration appears to be between 50 bp and 60 bp.  Treated as a rod, 60 bp DNA is 20 nm long.  The max pore size in the #101 is 14-15 nm.  If rotational diffusion of the rod like DNA is fast relative to translational diffusion, then one might expect a 15 nm pore to pass nothing bigger than 44 bp.  The data comes fairly close to meeting this expectation.  We believe that this cutoff will decrease further with decreased salt concentrations but it seems unlikely that is will increase with increased salt concentration.

We are gearing up for the next phase of testing membranes for DNA separation by diffusion.  A set of oligomers, that will form a custom made ladder, has been ordered.  There will only be 9 rungs in the ladder, from 8 bp to 224 bp, starting with a small step size of 8 bp and ending with a large step size of 64 bp.  We believe this will improve the accuracy of the measurements by gel electrophoresis.  Using membranes from a new wafer like #101, we are optimistic that reproducibility will get even better.

On a side note, our lab will be out of town for a stint.  Everyone will be back by June 17th.  I would like to request that we discuss this data on Monday, June 22nd if that would be all right with everyone.