Aloah,
I need your experience in the quantification of phosphorylated proteins via Western Blot. There are many papers with Western Blots and numbers for Quantification, but in most cases it's not written how they did. Actually, I doubt some results, since there are many techniques to do so, but the outcome differs quite a lot.
Using FIJI (or ImageJ), by drawing a box around one band or just the whole blot can make huge differences.
Now, I'm dealing with p-JNK, a target with two bands. However, using the antibodies for JNK and p-JNK (Santa Cruz, CST, respectively), I get three bands for all my tested human epithelial cells.
In the picture you can face many problems: The intensity of the bands differ quite well, so from the three bands in a lane, some are darker, for the next time-point lighter, but then another band is darker. Do you analyze them by phosphorylation site (for JNK it should be two) ?
The total protein of JNK seems less for p-46 than p-54, but their phosphorylation is strong at time-point two and three, how comes?
Another problem is the background, which is for some antibodies quite high, although I improved the protocol. When I want to quantify the band's intensity, I'll get a high value for the p-JNK at time-point one, since there's a shadow. But obviously, there's no real phosphorylation going on.
And: Reprobing might not the first choice for quantification (I've read). But if you don't have the fluorescence techniques for Western Blots, believe that putting controls or total proteins on another membrane is not very good, but having targets that are all the same size, how else should it be done?
So in general, I need your experience, not something I could google myself, which I did a lot, though. I know, there are many questions on RG, how to quantify protein phosphorylation via Western Blot, but nothing really answers my questions. Thanks for your help!
Tony
One thing you need is a loading control for each lane to normalize for differences in the amount of sample loaded in each well. It looks like you are using actin for this purpose.
You need to subtract the background. This should be done by defining an area near the band that represents the background in that part of the blot, if you are doing 2-D band integration.
If you use a 1-D quantitation, then you can define the background by skimming the base of the peaks.
I'd like to add one more point on top of Adam's suggestions: the traditional WB system utilizing a primary Ab and an anti-IgG-HRP has a drawback of non-specific band(s) around 40-65 kDa. We eliminated these non-specific bands by coupling HRP directly to the primary Ab, and got beautiful results. So, just conjugate the two Abs with HRP, and see what difference it will make. By the way, we use most Abs from Santa Cruz. Good luck.
Addition to Adam's suggestion, I use Image J to invert the WB results firstly. And then measure the signaling of those bands and the same size background region. The final value is got by using the mean value of your bands with the substraction of the mean value of background. Thanks.
Hi,
a valid normalization factor is very important, expecially after reprobing. Actin and other HKP are a questionable choice for many reasons. I would recommend a total protein labeling (e.g. Smart Protein Layers) to visualize the total protein at the two different target detection time points. This way you even calculate the protein loss into your normalization.
Programms like LabImage 1D (e.g. from NHDyeAGNOSTICS) reduce the background in each lane individually with a "rolling ball" tool. This makes background reduction very accurate.
Do you know how many variations your protein has? If you are interested in the regulation/phosphorylation of a specific form of your protein, a 2D-Blot might be helpful. The different phosphorylations on the different isoforms that are detected by your antibody are a problem. They even out the signal differences. Maybe you should have a deeper look into your protein. Maybe you are just interested in... let's say the membrane bound form but you see the one from the nucleus and the cytsolic variant as well. A pre-seperation might be an option here to get destinctive data.
For further questions about total protein labeling etc. please DM me.
Good luck,
Kristin
Your western blot bands look very sharp. Do you really want to distinguish these small variations of the phosphorylation?
If you just need to know the response of some treatment on p-SAPK, I guess your gel is too stiff.
Use 8-10 % PAGE, stop the electrophoresis halfway, and see the merged bold band, and quantify it by image J.
Repeat this several time, and put graphs of quantification of the bands. That's enough, isn't it?
Thanks so far for the replies. I'm looking forward for more experiences, so feel free to answer.
Tony
Dear Sei,
I am not sure if I understand your commend correctly but If you merge differentially regulated bands in one lane you loose the information about differentially regulated protein variants. And you even might loose a regulated signal in gerneral, if the different regulations eraze each other. If a protein has different variants, there is a biological reason behind a differential phosphorylation. I would want to see that. Even multible phosphorylations can be a relevant result. The main questions here are: How specific is my phospho-specific antibody and how can I normalize my signals? Using house keeper might not be a good choice in general. A number of biological replications have to be made in any case. A blot that only "looks" nice is not the goal of a Western blot.
Kind regards,
Kristin
Dear Kristin,
To distinguish these bands may become a hint of future study, but it depends on the purpose of the experiment.
The commercial antibody (Santa Cruz?) can't distinguish JNK3 from JNK1 clearly, JNK2 may play different way.
To see the p-SAPK is usually for seeing the cause of cell death or whatever.
In this context, JNK1/3 may play redundant roles each other. Increasing total amounts of whatever p-JNKs indicates that this treatment is going to kill the cells anyway, I thought it must be a purpose of this experiment.
As long as I know, if it is Cell Signaling Transduction's antibody against phosphorylated SAPK, the target of this antibody should be JNK1, it shows only 1 or 2 bands dependent on cell types. If he really interested in the order of activation of JNK1 and 3 (and the other), the resolution is still not enough to distinguish these, and most importantly, I don't think the "third" is differently regulated. In some condition, (sample preparation, gel stiffness, buffer condition etc) the phosphorylated band are delayed from inactive band, there would be an intermediate state of band. The relaxed condition of the gel may merge these small variations. That's what I'm thinking.
If you (Tony) really want to distinguish these, there is a way. Phos-Tag gel (Wako chemical industries, NARD institute, MANAC corp, I don't know the other countries' providers, sorry) separates the phosphorylated bands from the inactive band, this gel even separates by the degrees of the phosphorylation. Basically the phosphorylated protein binds to Phos-Tag reagent inside of the gel, and the artifactitious size of high-p-protein becomes bigger than the low-p-band. With the combination of specific RNAi or KO, you could determine which band is which JNK.
Best
I think the challenge for western blots is the quality of the primary antibody and protein degradation which occurs with time. In this case it is the anti-JNK antibody/anti phospho antibody. Since your question would be a phospho JNK related question and the western blot is a challenge, I would have tried a site-directed mutagenesis of the phosphorylation site , and observed for lack/inhibition of downstream signaling.
- Badges
- Science method