Filter Update – plan going forward

Since I’m now back pretty much full time on the filter project, I’ll give a quick update about where we stand and where I’m going with it, as well as some (negative) results from this week.

On the membrane transfer side of things, we finally have a tool that can produce fine enough vapor to do membrane transfer with something cleaner than my own breath. You can find it here. Since the mist it produces is cool, you need to either supply it with hot water or cool the chip to allow condensation, but it works well. Using that, I have about 90% yield in smooth membrane transfers, then 3 PDMS painting steps with about 90-95% yield each, which means I get transferred, PDMS-stabilized membranes about 70-80% of the time.

If you recall from the last update at the end of last year, we are at a point where we can reliably make pores in filter systems, but we were having a hard time getting any DNA to actually pass the membrane. We had one experiment which shows translocations, but it was not reproducible. The plan now is to use the filters as a proof of principle for an assay that uses very short (50bp) DNA, which we hope will be easier to capture going forward.

The assay is based on cleaving 50bp dsDNA from large gold nanoparticles to which it is initially hybridized. The DNA could then pass the membrane, while the particles would be excluded by the filter. The hope is to show that without the filter, the gold nanoparticles will quickly clog the pore. We have some preliminary data on regular membranes showing that this is the case, so the task now is to demonstrate the same thing using the oxide-covered chips.

The plan looks like this:

1. Show that the patterned oxide chip (no filter) will quickly clog in the presence of 50-bp dsDNA-coated 50-nm AuNPs
2. Show that the patterned oxide chip (no filter) will show translocations of 50-bp dsDNA
3. Show that the patterned oxide chip (no filter) will show some translocations then quickly will quickly clog in the presence of 50-bp dsDNA-coated 50-nm AuNPs mixed with free 50-bp dsDNA
4. Show that the patterned oxide chip (with filter) will never clog in the presence of 50-bp dsDNA-coated 50-nm AuNPs
5. Show that the patterned oxide chip (with filter) will show translocations of 50-bp dsDNA
6. Show that the patterned oxide chip (with filter) will show translocations of 50-bp dsDNA without clogging in the presence of 50-bp dsDNA-coated 50-nm AuNPs mixed with free 50-bp dsDNA

If we can get all that done, we have a pretty decent paper.

This week I started with experiment 1, with no success so far. The first chip I mounted had some ragged edges that caused it to shatter in my flow cell. I was able to make a nice pore in the second chip, but the cell had a very slow leak (I think the PDMS layer may cause poor sealing with the gaskets, and may require thicker gaskets in the future). I was able to salvage that and stop the leak, but adding gold nanoparticles with DNA showed no signals or clogging.

There are a few possible reasons for this: one is that the gold particles are not properly hybridized. We are running a kit to make them ourselves, and it seems to be working, but it’s difficult to say with certainty. Secondly, in order to the nanoparticle suspension to be stable, I have to work in very low salt concentrations. Low salt is not a regime we have really explored before, so it’s possible there are some traps there that we are unaware of as of yet. The third possibility is that the 1-micron oxide holes are sufficient to exclude nanoparticles from the nanopore region. This seems less likely because of the holes being so shallow, but may be worth considering.

Next week, the plan is to rule out the third option by using the same sample in a regular nanopore membrane without the oxide. If no signals are seen, we will conclude that the nanoparticle batch is to blame and start over (we already have some data on regular membranes with an older batch of AuNPs showing the expected clogging). I will also start to run 50bp dsDNA through failed AuNP experiment pores just to see if the DNA can be seen at all. Once the initial issues are debugged, I’ll move on to using actual filters.

Let me know if you have comments or objections to the plan going forward, or if there are other experiments you would like to see.