You’re gathering data from all your experiments and preparing to present to your advisor and thesis committee at your annual progress report. You have an interesting hypothesis, and you have a validated antibody that recognizes your target protein on a western blot (WB). The molecular weight of the band is correct, and the expression of the target protein changes just the way you predicted it would. Now, you know — and you’d bet the house on it — when that powerpoint slide comes up, someone on your committee is going to ask about loading controls.
Whether setting up a new batch of WB experiments, or interpreting WB data in the literature, there are two questions that should be in mind when comparing WB bands: 1) Are the lanes loaded equally? 2) Is the signal in the linear range of detection?
To compensate for sample variation and ensure equal loading, cellular protein content is frequently measured in the form of loading controls/“housekeeping” proteins such as β-Actin or GAPDH (see “How Do I Choose A Loading Control?). In theory, the quantity of housekeeping protein on a blot will be proportional to cell number, but this is an assumption that can be misleading, especially when a single protein is used for normalization (1). While it is commonplace to use a single loading control, increasing the number will lead to lower variance. The use of 5 loading controls is considered to be optimal (2, 3).
Unfortunately, loading control proteins are often far more abundant than the protein of interest. When the control and target protein are measured at a single dilution, the signal for the more abundant protein may be saturated, exceeding the linear dynamic range of detection. Saturated control bands under-report variance between lanes. You can probably appreciate how this problem could reinforce assumptions about equal loading, and possibly lead to spurious conclusions about the experimental variable!
Sometimes, overloading and saturation will lead to misshapen blobs instead of orderly bands, or bands with “burned out” or “hollowed” centers indicating that the HRP-conjugated secondary has exhausted the local concentration of the chemiluminescent reagent. This may be accompanied by brown spots on the membrane, a byproduct of peroxidase hyperactivity. Shown below is an extreme example of this phenomenon.
Ever get that burned-out feeling?
However, note that the absence of “weird” bands or brown spots on your blot may not be indicative of linear signal. Even if the control bands appear well-ordered, they could still be saturated (2). To demonstrate linearity for both the loading control(s) and protein(s) of interest, load a range of sample dilutions on the same blot, or multiple blots transferred simultaneously to avoid variations in transfer efficiency (2, 3). Titrating down the protein input, the primary antibodies, and/or chemiluminescent reagent may be necessary to achieve linearity, particularly for highly-expressed loading controls. Below is an example where acquiring a longer and a shorter exposure was sufficient to achieve linear signal for HNF1 and GAPDH, respectively.
WB signal is also highly dependent on the detection method. Enhanced chemiluminescence (ECL) detection by a charge-coupled device (CCD) camera results in superior linear dynamic range compared to film. Software settings can be employed to display red pixels if the digital signal is saturated. Fluorescent scanning (using secondaries conjugated to fluorescent dyes instead of peroxidase enzymes) may prove better still by avoiding the limitations inherent in the chemistry of luminescence.
These approaches can mean fewer serial dilutions when setting up your experiment. But if your lab doesn’t have access to the latest and greatest technology, you can still get good data using traditional ECL and film, provided you take the time to set up the experiment to address the equal loading and dynamic range issues.
All that said, using multiple loading controls is obviously a time- and reagent-consuming issue. An increasingly viable alternative method is to measure total protein loading in each sample lane, a logical extension of the multiple control concept that avoids the need to probe for multiple loading controls. Measuring total protein involves staining and scanning all the protein bands in each lane (1), or at least multiple discrete bands (3). It is prudent to either confirm the stain does not interfere with subsequent immunodetection (potentially dependent on the antibody), or run duplicate blots, one for total protein and one for the antibody.
For many grad students and postdocs, running western blot experiments might feel “routine”. But it’s important to remember that when interpreting data from WB experiments (as it is for any experiment), the devil is in the details. Careful consideration of issues such as sample preparation, transfer conditions, and primary antibody, among others, will help you avoid pitfalls (2, 4-5).
And when the time comes to publish, remember to provide all the details to allow others to accurately assess the results and repeat the experiment. In other words, blot responsibly!
- Eaton SL, Roche SL, Llavero Hurtado M, Oldknow KJ, Farquharson C, Gillingwater TH, Wishart TM (2013) Total protein analysis as a reliable loading control for quantitative fluorescent Western blotting. PLos One 8(8):e72457.
- Janes KA (2015) An analysis of critical factors for quantitative immunoblotting. Sci. Signal. 8(371): rs2.
- McDonough AA, Veiras LC, Minas JN, Ralph DL (2015) Considerations when quantitating protein abundance by immunoblot. Am J Physiol Cell Physiol. 308(6):C426-33.
- Ghosh R, Gilda JE, Gomes AV (2014) The necessity of and strategies for improving confidence in the accuracy of western blots. Expert Rev Proteomics. 11(5):549-60.
- Gorr TA, Vogel J. (2015) Western blotting revisited: critical perusal of underappreciated technical issues. Proteomics Clin Appl. 9(3-4):396-405.