Here is a dilemma: there is a line of null mice (global knockout), which are null by genotyping and by mRNA analysis (qPCR). They were generated by replacement of an entire protein-coding region with a Neo cassette using homologous recombination. However, commercial antibodies raised to the N-terminus of this protein still detect it (on western blot). The signal is comparable to that in wild-type mice, with regards to the protein size (kDa) and band intensity. Thoughts?
Talked to a few people in my lab and they all said the same thing. Did you run a secondary only control? Because, as long you are confident in the genotyping, it sounds like the antibody is picking up a non-specific band in both samples. If the secondary only control is blank. Then the antibody is not against the protein that has been ablated in your mouse line. Probably grounds for compensation haha.
I agree with with Johnathan. It may just happen that your antibody cross-reacted with a non-specific protein of the same size on the gel. If you confirmed the knock-out by genotyping and qPCR analysis of the corresponding mRNA, it would be impossible for the cells to produce the protein. So, is there a homologous gene in the genome? Can you detect a homologous sequence by blast search with your gene/protein of interest, especially against the N-terminus of your protein against which the antibody was generated? Also you may want to try different dilutions of your antibody? Maybe a non-specific protein of the same size, but is just so abundant? Good luck with your experiments.
Hi Johnathan and Hua,
Thank you very much for your insights and helpful questions. I greatly appreciate it! I am considering a possibility that secondary antibody alone might be picking up non-specific bands. Will be running more gels to test this idea. Also, the antibodies I've tried so far were raised against the N-terminus of the protein. This week I will test an antibody against the C-terminus of the protein. Meanwhile, I will definitely be checking for any homologous genes and sequences in the N-terminus. I will be contacting the companies to get help with that, since the sequences they use to generate antibodies are proprietary.
Sure, the antibodies you used are not specific. For references see:
https://www.nature.com/news/reproducibility-crisis-blame-it-on-the-antibodies-1.17586
Article Editorial: Antibody Can Get It Right: Confronting Problems o...
https://assets.thermofisher.com/TFS-Assets/BID/Reference-Materials/antibody-reproducibility-validation-specificity-tests-nature-articles.pdf
https://www.nature.com/articles/529025c
As said, your antibodiea are not specific for your target protein.
In my opinion, you have to change the antobodies (source!) and test new antibodies before using in experiments (esp. via recomb. protein itself - would show specificity to your target protein -).
Have you find the problem? We have exactly the same situation currently. We generated a global knockout mouse using CRISPR, the PCR showed knockout, but western blot detected the same band(similar kDa) compared to control mice. Interestingly, we also found that cellular function(A protein changed in response to B knockout) was changed in the knockout mice. So I will take the suggestions above, and see if it is the problem of antibody. Thanks for your suggestions!
Liang,
Yes, I found the problem. The protein lysates I am using are from mice that have a lot of inflammation in the CNS. I discovered that non-specific bands in the knockout were from mouse secondary antibody recognizing light chains on the blot (at ~25 kDa). By coincidence, the protein I am analyzing also blots at ~25 kDa. There are ways to circumvent this problem, which I am pursuing right now.
- Badges
- Science method