BSA adsorption comparisons
I set out to do an in depth study of adsorption to various surfaces using the sample buffer method. Midway through staining the multiple gels I noticed there was a problem. It looked as though glass and polystyrene adsorbed about the same amount of BSA as did pnc-Si. Details of the setup follow.
Preparation:
Using vaccuum grease I outlined small circles on glass coverslips and untreated polystyrene culture dishes. These circles were about the same size as our pnc-Si chips, as measured roughly with a micrometer. I then applied 30uL of BSA to 5 pnc-Si chips, 5 glass coverslip circles, and 5 polystyrene circles. The BSA was removed after 30min and the samples were rinsed with H2O by pipetting on 30uL 2x and pipetting up and down to gently agitate. 30uL of 1x sample buffer was added to the samples to removed the adsorbed BSA. The sample buffer was removed after 30min and the samples were prepared for SDS-PAGE along with 3 mass standards in the linear range (described in my previous posts).
Results:
The 10% SDS-PAGE gels were silver stained and digitized. Using the ImageJ densitometry method I measured the band intensities and calculated the concentrations by fitting a line to the mass standards. MatLab was used to collect the data and plot in the following figure.
Notes: pnc-Si samples from w219, non oxidized, pinholes
Updates: Figure changed to show protein mass adsorbed to surface.
Here is one of the many gels stained for this study:
Your studies, taken together, would suggest that PES and cellulose bind to protein as much as glass or polystyrene surfaces. That doesn’t seem right to me. On the other hand, why should our material be less sticky than glass? I thought we expected this because pnc-Si was atomically smooth and glass isn’t, but we saw pictures a few weeks back from the new STEM that suggested pnc-Si is not very smooth. I really don’t know where to place my expectations, but your old fluorescence studies will need to be redone because we clearly are not getting the same answers as we did then.
To believe this assay is really working, we’ll need to show that we can get a significantly different number with some surface. Looks like PEG silane (should bind less) or amino silanes (should bind more) might be the next step? Maybe teflon?
Also note that your vertical axis doesn’t really make sense. When you quantify a gel, you load known volumes in each lane and so you are really quantifying the mass of protein in each lane. Thus you should be reporting ug protein here instead of ug/ml.
Even better is to divide by the area and make this mass/unit of material surface area. If you do that we can think about the reasonableness of your numbers from a theoretical perspective using albumin size and parking arrangements. Ask Mort for help with that.
I believe that there is a problem with this method. Even if our membranes bind as much protein as the glass, I’d expect that the other treated membranes would be significantly less. I can attempt the fluorescent studies again the week.
I’ve changed the figure to represent mass of protein adsorbed. The amount of surface area is held to be about the same in all of these studies. For pnc-Si membranes it looks like there’s .0075ug for each square mm.
It’s a shame that we have not had a meeting in a while, there were a few issues that I wanted to raise that I thought were best to save for a discussion. I have two main points:
1) Why rinse with DI water? This is something of a red flag to me, as many proteins behave in unpredictable ways when they are removed from buffer. The dynamics of how the buffer is rinsed away by DI also come into play at the surface, and I would expect considerable variability. I would try again with a buffer rinse instead. Also, if at any point the surface dries, this will have a strong effect on your results.
2) What are you really measuring? I think you are measuring protein that is bound strong enough to stay in place during a short gentle rinse, but week enough to be soluabilized in 30 minutes in buffer. An unknown amount of protein remains on the surface, so you are not measuring the total bound protein, but only the weekly bound fraction. This could be useful, but it seems like that total bound protein is the number that is most relevant.
If you are interested in “seeing” what happens on oxide surfaces during your protocol, I could give you a few Pathologics chips to play with and then we could image them. It will be difficult to quantify, but you would know what conditions leave the least amount of protein behind. We detect down to a few pg/mm^2.
BTW-7.5ng/mm^2 is pretty close to a dense monolayer of BSA, depending on the orientation that you use.
This assay is used all the time with nanoparticles, although we are the only folks I’ve seen use it in a quantitative way. Still we are accustom to reproducible results and significant differences between materials, kinetic adsorption , etc, that make sense. If this is a monolayer of BSA, the numbers here look reasonable once again. Still, we need a known low-stick material to give us faith in what we are doing. I like dense PEGs on glass (or pnc-Si) since we’ve had our eye on it for a long time.
The ‘buffer’ that removes the protein from the surface is Laemmli ‘sample buffer.’ It was designed to aggressively denature and solubilize proteins into individual peptides for gel electrophoresis and so we assume no protein would be adsorbed strong enough to a surface to survive. Since a few protein-protein associations are known to survive under these conditions, its conceivable that some protein-surface interactions might too. Still, I bet we remove >99% of the protein. It would be nice to validate this assumption, but the approach would need to be able to tell us we are back at a signal from a clean surface after the SB wash. If the pathologics system can do that, this might be a starter project for Rachel.
Jess – For breaking down protein-protein interactions the buffer is normally boiled first. I’m not sure if you are doing that. If not, you might want to throw that step in with a PBS wash version (addressing Chris’ point) of this assay.
Once again it seems we are diving deep to answer what was supposed to be a simple question, but getting this right will be quite valuable for you and the whole project and we are still too low on materials to wrap up the movable cut-off studies.
I will try the buffer rinse this week to see if it makes a difference. I’ll also try to set up some PEG surfaces.
Jim, I’m boiling the buffer before running it on the gel. I’m not sure what you’re suggesting here. Do you mean to boil it while on the chip, or to boil it before adding to the chip?
Yes. I’m suggesting that you add hot sample buffer. Do you see any problem with that? Its probably not going to make a difference, but if you are redoing these experiments, you should use the most aggresive conditions.
What’s in this special cleaning buffer? Is it an enzyme of some kind? Denaturing typically causes more binding to surfaces, in my experience, but if you are also chopping it up into pieces, I can see it becoming soluble. As long as the surface remains wet, you may be able to get most of the protein off – maybe. I’ve exposed to chips with dried monolayers of protein to concentrated acid, base, and various organic solvents, with very little sign of removal.
Another thought – you may also want to pre-wet your surface with clean buffer, before adding your protein solution. The interaction of a dry surface with a protein solution is complex, because there will be a skin layer of protein on the air-fluid interface of the solution. In the nanoparticle assays, I doubt there is ever a “wetting” of a dry surface, and this could dominate your current situation.