n-Dodecyl β-D-Maltoside Coating on PDMS

Today I performed a coating of n-Dodecyl β-D-Maltoside (“DDM”) on PDMS in an attempt to block protein adsorption (reference). The reference describes two different procedures, one in which the protein being incubated on the PDMS surface is in solution with DDM, and one in which the DDM coating is performed before the sample is washed with buffer followed by incubation with protein only. For these purposes, I’ve chosen the latter, in spite of its slightly decreased effectiveness, in order to fill the needs of the device as a product to be sold.

I prepared what the reference refers to as “DHBS,” which is 0.1%-by-mass DDM in a Hepes buffer. The Hepes buffer (“HBS”) was prepared beforehand in accordance with this recipe and frozen for storage. I unfroze the solution and mixed it in a 1:1 ratio with DI water to dilute it to a 1x solution. I then weighed the 1x HBS and added one one-thousandth that mass of DDM to it.

Eight strips of PDMS were prepared: two treated with DDM (1, 2), two “sham treated,” or given the same procedure but with simple HBS instead of DHBS (3, 4), two completely untreated with buffer (5, 6), and two untreated and without any protein to establish a background fluorescence value (7, 8). Samples 1 and 2 were placed in a petri dish and bathed in DHBS for five minutes. Samples 3 and 4 were also placed in a petri dish and bathed with simple HBS for the same durration. Samples 5 and 6 were placed in a petri dish with no bath. 50uL of 1mg/mL fluorescein conjugate BSA was pipetted onto samples 1 through 6, and left to incubate for one hour with all petri dishes shut and covered by tin foil. Following incubation, samples 1 through 6 were washed twice with simple 1x PBS, which was added by pipette directly onto the samples to the point of submerging them, decanted, added again, and decanted again. These samples were then placed, covered loosely by tin foil, on a hotplate at 100 degrees celsius to accelerate drying.

Fluorescence measurements were taken with FITC excitation on the fluorescence microscope, and the results of analysis by Imagej were as follows:

	Fluorescence	Treatment
Sample 1	114	DDM
2	151	DDM
3	154	Buffer only
4	156	Buffer only
5	110	Untreated
6	115	Untreated
7	64.0	No protein
8	66.6	No protein

These results are almost extremely clear in their meaning, but unfortunately, sample 2 muddies things up by displaying fluorescence to be expected of buffer-only treatment in spite of the DDM it was exposed to. In order to clear this up, I plan on doing more attempts at DDM coating in order to determine which number — the 114 or the 151 — was a fluke (fingers crossed!)

Something that the reference paper mentions as being important to this process is “continuous washing,” which is described as “flowing buffer at 1 mL/hr for 3 min (50 uL).” I would like to copy this procedure exactly if possible, but I was unable to determine how to get it to happen for a simple strip of PDMS rather than the PDMS channel that was used in the paper. Jamie and I came to the conclusion that simply submerging the samples in excess buffer might accomplish the same thing, but it’s conceivable that the mechanism by which the buffer washing helps to stabilize the DDM coating relies on the constant presence of fresh buffer near the coated surface. If anyone has any ideas about this, please let me know, as I would greatly appreciate it!

As an aside, when I was performing the DHBS bath I noticed a strange phenomenon in the petri dishes that I’m unsure what to make of. I took a picture of it:

This fogginess had a sort of rainbow effect to it and was on the inside of the petri dishes, on the bottom. It occurred in both DHBS-treated samples (left) and HBS-treated samples (right) but was much more pronounced in the DHBS case, which made me suspect that the DHBS was reacting in an unexpected way. I don’t know if it had any effect on the experiment or if I’m just imagining things, but if someone has seen this before and knows what it is, I’m all ears. I’ll certainly be keeping an eye out for this sort of effect in future coating attempts.

UPDATE: I’ve just finished with three more DDM-coated samples using the same procedure. The results were as follows:

	Fluorescence	Treatment
Sample 9	89	DDM
10	95	DDM
11	138	DDM
12	46.8	No protein

The background displays roughly 20 units less fluorescence today, so in order to compare these values to the previous values, it seems reasonable to add 20 to each. Samples 9 and 10, then, are comparable in fluorescence to sample 1 from yesterday, while sample 11 is comparable to sample 2 (as well as 3 and 4.)

I have a convenient explanation for this today: I didn’t prepare quite enough DHBS, and sample 3, being that last one, didn’t get quite as much expose to it — about 8mL instead of 10mL. This made submerging the sample in the DHBS much more difficult, and could explain the poor coating.

It seems fairly clear, then, that proper DDM treatment results in significantly decreased protein adsorption. As soon as possible, I’ll move on to performing similar fluorescence measurements on PEGylated glass and SiN samples, before proceeding to mortar bond them all together into a CytoVu assembly and perform simple separation experiments.

UPDATE 2: Jim and I discussed these results and it was decided that better data might be taken if the samples were placed on the opposite side of the glass slide on the fluorescence microscope. Previously, I had taken measurements with FITC excitation, where the excitation laser passed through the glass slide, through the PDMS, and then the emission light took that same path back down to the camera. I repeated the process with the PDMS stuck to the bottom of the glass slide instead, and took fluorescence measurements where the excitation laser passed through an 0.6 filter to cut back on noise:

	Fluorescence	Treatment
Sample 13	73.2	DDM
14	131	DDM
15	90.9	Sham
16	105	Sham
17	53.2	No protein

Unfortunately, these values are no longer comparable to the values for samples 1-12, since they were not measured in exactly the same way. Again, the DDM samples display wild variation, this time in fact including one sample which showed more fluorescence than sham treated samples.

As a possible solution to this (and for general edification,) Jim and I also discussed increased the concentration of DDM in the HBS and seeing if a more even and thorough coating could be achieved. It’s very possible that the variation in fluorescence observed in DDM treated samples could be due to insufficient DDM for an even coating across the entire PDMS surface. The reference uses a very small surface area of PDMS (only 5.2 square microns) whereas my samples are roughly 4 square centimeters, as well as 1mm thick. Thus, I plan to spend tomorrow testing increased concentrations of DDM. I’ll update this post again when the results of that venture are in.

UPDATE 3: I performed DDM treatments and fluorescence measurements using 1%-by-mass DHBS. This represents a ten fold increase in DDM. I also increased the incubation time of the DDM on the PDMS surface, from five minutes to half an hour, in order to give the coating more time to reach equilibrium between the surface and the solution. Fluorescence measurements were taken in the same manner as the measurements taken for update 2. Results are, unfortunately, ridiculous:

	Fluorescence	Treatment
Sample 18	75.6	1% DDM
19	123	1% DDM
20	111	1% DDM
21	88.7	Sham
22	92.6	Sham
23	53.4	No protein

Not only do DDM-treated samples continue to display huge variations in fluorescence, but this time two of them displayed more fluorescence than sham-treated samples, indicating that either an excess of DDM leads to a more adsorptive surface, or my measurement techniques are flawed. I cannot rationalize any way in which the chemistry could lead to so much more protein adsorption, so I’ll be looking into ensuring my techniques are reasonable by incubating the same volume of various concentrations of fluorescent BSA on a glass slide and taking measurements. If they prove sensical and consistent, then I’ll have to look into other possible causes of these absurd discrepancies. Perhaps sonicating samples before coating will help to make things more reasonable.

UPDATE 4: I’ve performed benchmark tests by incubating various concentrations of fluorescent BSA on a glass slide. I made eight 50uL pools in total, two of each of the following: 100% BSA, 7:3 BSA:PBS, 3:7 BSA:PBS, and 100% PBS. One of the 7:3 pools mixed with one of the 3:7 pools, contaminating those measurements, but that’s why I did two of each in the first place.

I won’t bother posting exact measurements here unless they’re for some reason desired, suffice to say that they behaved precisely as expected. That’s encouraging, because it removes the variable of faulty measurement technique from the equation. I’ll be redesigning my experiments to further remove variables, I think by making the PDMS samples smaller so that the are covered by the 50uL pool of BSA represents more of the surface area of the sample — ideally, all of it. This should ensure that I’m placing the BSA in the same relative location on each sample, and consistently measuring in that location as well. It should also reduce the amount of DDM required to create an even coating on the PDMS surface, approaching the size of the PDMS surface covered in the reference.

This post is becoming extremely long, so this will be my last update to it. Future updates on this topic will be in a new post, and hopefully will be more interesting due to more sensical data!