Editor’s note: since the original publishing date of this article, AzureSpot software has been updated to AzureSpot Pro. Learn more about AzureSpot Pro here. Try a free 2-week trial!
Best practices for quantitation of unphosphorylated vs phosphorylated forms
In part 1, we reviewed tips for using multiplexing to visualize phosphoproteins. In part 2, we’re going to cover quantitation. Once you have reliable, repeated samples that fluoresce in non-overlapping bands and are linearly proportional to the signal, you can begin analysis. Here we will use the AzureSpot analysis software.
Background Subtraction
Background subtraction is essential for effective quantitation. This is where analyzation software comes in handy. AzureSpot, for example, has five automatic background options and three manual options.
Automatic options
Figure 2. Background options on AzureSpot
Figure 3. Examples of results from two different sizes of rolling ball for the same blot. Purple line is the background
Use Rolling Ball if you have an uneven background. A disc with a specified radius ‘rolls’ under the lane profile, averaging the signal, which smooths out small variations. The larger the radius, the more smoothly the disc rolls under the profile.
Use Rubber Band if the ends of the profile are not too different from the rest of the profile. This method works as if you are stretching a rubber band under the lane profile.
Be careful of using this method where the lowest point is found at the edges of a lane because you might draw the lane slightly differently later and the values will be different. Generally, this calculation is hard to repeat from analysis to analysis.
If you’ve already performed band detection, you can use Valley to Valley, which uses the edges of each band to determine and subtract background. It’s important to enter a maximum slope to avoid identifying the edge between overlapping bands as background.
Minimum Profile sets the background as the lowest value found in the profile.
Lane Edge Subtraction uses the lower of the two values at the edges of each lane. Only use this method if the bands are completely enclosed because if the edge is drawn on a bend, it can skew the results.
Three manual options:
- Image Rectangle finds the average intensity within a rectangle drawn on the image. In a multiplex gel, the same rectangular region is used across all of the channels, though slightly different values specific to those channels are set as the background for that channel.
- For Image Stripe, draw a thin, vertical rectangle alongside the lanes. The average intensity at that point is the background for that point in the lane. Again, the region you draw is used for all channels, though the specific values will be different, specific to each channel.
- For Manual Profile Subtraction, determine the shape of the background in theprofile Analysis window. For multiplexing, you can edit the background for each channel separately. The manual baseline method subtracts the background in the selected lane only, and recalculates the background every time you make change in the Analysis window.
Normalization
AzureSpot can perform normalization in three ways. First, it can set the normalized volume for a specific band in a lane in a standard gel. In a multiplex gel, it can normalize with housekeeping proteins or total protein.
Total protein normalization (TPN) is becoming increasingly common because of its greater accuracy and suitability to quantitative western blotting. If you have validated that the expression of your housekeeping protein doesn’t change, it might be suitable. Otherwise, TPN is preferred. Even the most reliable housekeeping protein method can’t account for factors like membrane transfer variability, which TPN can.
Treat the membrane with a total protein stain (e.g. AzureRed Fluorescent Total Protein Stain) after transfer but before immunodetection. This method is antibody-independent, and accounts for variation in sample protein loading and transfer efficiency.
Don’t forget to use a dilution series to confirm that you’re working within the linear range of detection, which ensures that the signal intensity is proportional to the amount of sample (as discussed in part 1).
Be careful to load the same amount of sample protein across the entire gel. A protein concentration assay could help determination if an adjustment in concentration is needed.
Make sure you’re in the linear dynamic range before normalizing because this range varies by concentration.
- AzureSpot performs normalization automatically. The purple line will always be the background, and the green line will be the signal. First select a lane to be the reference lane (or normalization standard). For each lane, calculate the normalization factor, volume of all material in the lane divided by all material in the reference lane. Next calculate the normalized volume of the bands in both channels, which is the raw volume of the band divided by the normalization factor.
Figure 7. Normalization and signal curves in AzureSpot
Even though AzureSpot will make these calculations automatically, here is a brief review of the procedure:
Store the total protein staining and target protein quantification data in an array.
Calculate the mean, standard deviation, and coefficient of variation of each sample using the total protein stain data from the normalization channel,
Remember that coefficient of variation is the ratio of standard deviation and mean.
Interpreting the Results
Use the normalized Target Protein values to determine the relative amounts of each sample. Coefficient of variation can help determine how reliable the results are. If the CV value is low, then the signal variability was low, and precision is higher. A larger CV means that there was greater variation in the signal and the measurements were less precise.
Consider a difference in band intensity to be a reliable result if it is greater than the CV. There are different standards, but a good rule of thumb for many cases is that if the magnitude of the difference should be about double the CV, or two standard deviations above the mean.
Compare the effect that normalization has had on the CV. CV should be more or less the same before and after normalization. If normalization causes a large increase in CV, then there may have been an error in the experiment.
With effective background subtraction and total protein normalization, you’ll be able to take multiplex images of phopsho-proteins and perform reliable and effective quantitative analysis.
Next calculate the Lane Normalization Factor (LNF). To this, take the lane with the highest signal for total protein staining (TPS), and then:
Next take the target protein data, and repeat. Calculate mean, standard deviation, and coefficient of variation.
Calculating the Normalized Target Signal for each band requires the LNF.
Remember that normalization is specific to only this blot, and must be repeated each time.
Interpreting Results
Use the normalized Target Protein values to determine the relative amounts of each sample. Coefficient of variation can help determine how reliable the results are. If the CV value is low, then the signal variability was low, and precision is higher. A larger CV means that there was greater variation in the signal and the measurements were less precise.
Consider a difference in band intensity to be a reliable result if it is greater than the CV. There are different standards, but a good rule of thumb for many cases is that if the magnitude of the difference should be about double the CV, or two standard deviations above the mean.
Compare the effect that normalization has had on the CV. CV should be more or less the same before and after normalization. If normalization causes a large increase in CV, then there may have been an error in the experiment.
With effective background subtraction and total protein normalization, you’ll be able to take multiplex images of phopsho-proteins and perform reliable and effective quantitative analysis.
Figure 8. AzureSpot automatically detected lanes
How can I learn more?
Learn more about AzureSpot here and download a free 2-week trial!
- Ghosh, R., Gilda, J.E., Gomes, A.V. (2014). The necessity of an strategies for improving confidence in the accuracy of western blots. Expert Rev Proteomics. 11(5), 549–560
