Validating your antibodies, i.e. confirming that an antibody recognizes your protein of interest with low cross-reactivity to other targets, is critical for ensuring consistent, reproducible results. Validation is frequently required before submission to a journal, but not everyone takes the time to validate. Or many of us assume that because we validated the antibody for immunopreciptation, we don’t need to validate it for western blotting. However, binding is highly dependent on experimental conditions, so it’s important to validate antibodies in similar conditions to those in your protocol.
Quantitative western blotting in particular depends on confirming that the signal you’re getting is coming only from your protein of interest. Without antibody validation (and validation that the signal is linearly related to the amount of protein), you won’t have the accuracy and precision needed for quantitative analysis. After all, the goal of the study is to measure the abundance of a specific protein, and cross-reactivity can artificially inflate your abundance measurement.
To ensure experimental quality, the International Working Group for Antibody Validation gives five different methods for validating antibodies, four of which work for Western Blotting, and they recommend using at least two of these methods .
This approach should be familiar to every biochemist who’s been in the game for a while, though the methods have been updated! Measure the relevant signal in control cells or tissues in which the target epitope has been knocked out or knocked down using CRISPR-Cas9 or RNA interference (RNAi). Because there should be very little to no expression of the target protein, any signal observed will indicate crossreactivity.
In this western blot, we see protein lysate from a cell line with a particular protein knocked down by two different siRNA molecules in columns 1 and 2. Column C is control. It’s possible to determine how much off-target binding there will be by comparing the columns.
This next method is a little more low-tech. Use an antibody-independent method to quantify the target signal for several samples and then compare these amounts with those from antibody-based approaches. For example, there are several targeted proteomics approaches that can quantify protein expression in different samples. Antibody labeling should correlate with these quantifications.
This figure compares quantification of protein expression in eight different cell lines using western blot and mass spectrometry. With a sufficiently high correlation, we can conclude that the antibody is valid for western blot analysis.
Independent antibody strategies
Using two or more different antibodies against the same target is an intuitive and effective method to gauge the specificity of an antibody. It’s important to test the correlation within one application environment. The antibodies are independent if they have different epitopes (bind to different regions of the protein). That way they are unlikely to exhibit off-target binding to the same protein that’s unrelated to the study. This method pairs well with a knockout or knockdown approach.
Two independent antibodies are used against a protein, and we can determine the correlation of the signals from each of the eight cell lines. Be sure to check the entire gel because bands indicative of off-target binding may be elsewhere.
Expression of tagged proteins
It is also possible to validate antibodies by expressing a protein that has an affinity tag (e.g. FLAG or v5) or a fluorescent protein (such as GFP or YFP). It is then possible to match the antibody expression with that of the other method. This approach is limited because it’s best to tag the endogenous gene. Overexpressing the target protein might artificially drown out off-target binding, and then the method won’t be effective for validation.
Not only is it important to validate antibodies for Western Blot, but Western Blot is used to validate other methods such as ELISA and IHC.
Uhlen, et al. A proposal for validation of antibodies. Nature Methods 13(10), 2016, 823-827