Housekeeping genes/proteins, internal references/loading controls & related terms… “All” cells in your body have the ability to make every protein you’d ever need. But just like you don’t need snow boots in the summer, and you never need them in Southern California, your cells don’t need all those proteins all the time, and different cell types have different needs that can change depending on the time and environment. Bottom line, the levels of most of these proteins will vary between cells and under different conditions. We can say they’re differentially expressed (more on this all in a second). And if we measure this expression (such as with qPCR or western blots) we can learn about what’s going on in the cells. But in order to fairly compare expression between samples we need to normalize them – remove any bias that comes from having a greater total starting amount of sample. Otherwise you won’t know, for instance, if you’re seeing a stronger signal because the protein you’re looking for really is expressed more or if you just started with more cells. In order to do this normalization, we often turn to housekeeping genes/proteins to serve as internal references or loading controls.
Housekeeping genes/proteins are more like underwear than snowboots. They’re something you’re always needing and thus always making no matter where you are or what time it is (showers and nudists excluded for the sake of the analogy). Common “cellular underwear” we use as housekeeping genes/proteins are GAPDH (an enzyme involved in helping break down sugar in glycolysis), actin, & tubulin (components of the cytoskeleton – a network of tubes and stuff that gives cells their shape and moves stuff around in them).
In jargon terms we can describe housekeeping genes as things that are constitutively and ubiquitously expressed. Constitutive refers to something that’s going on “all the time” without needing to be told to (“no” regulation of activation). And ubiquitous refers to it happening “everywhere” (“all” cell types). We can use these terms in various contexts. For example we can talk about some enzymes and receptors having constitutive activity if they don’t need to be stimulated in order to be active). But, for housekeeping genes/proteins, the “something happening” that we are referring to is gene expression. Basically the genes getting “used” (transcribed into RNA which can then be translated into protein).
Much much more on gene expression as well as how we can measure it in other posts, especially: https://bit.ly/expression_measurement; YouTube: https://youtu.be/aTbJAnF4nqI
Since this is a more technical post, I’m going to assume people have a grasp of the basic central dogma – that is:
- the permanent recipes for making proteins are written in the form of segments of DNA called genes
- Messenger RNA (mRNA) copies of those recipes get made in a process called transcription
- Protein gets made following those recipes in a process called translation
There are various forms of regulation that can occur at each of these steps, which allows cells to control what they make, how much they make, and when they make it. This lets your cells differentiate themselves from other cell types and respond to changing needs.
By measuring the levels of mRNA transcripts (such as with qPCR) and/or protein (such as with western blots) we can get a glimpse at how highly various genes of interest are being expressed. And the real power comes from being able to compare this expression under different conditions. But that power relies on a way to normalize your results. Enter the housekeeping genes/proteins.
More details in the video, but basically you measure levels of a housekeeping gene or protein (e.g. tubulin, actin) in the same sample as the thing you really care about, then divide by the relative amount of the housekeeping protein to control for differences in how much sample you started with.
So, for example, to normalize western blots using housekeeping proteins as loading controls, the basic procedure is:
- quantify band intensities (such as with imageJ or similar)
- find the lane with the most housekeeping protein
- divide the amount of housekeeping protein in each lane by that highest amount to get the lane normalization factor (LNF)
- divide the band intensity for the protein of interest in each lane by that lane’s LNF
If you’re wondering how you can probe for different things on the same blot, fear not! I have a whole post on that: blog form: https://bit.ly/western_reprobe ; YouTube: https://youtu.be/XUb-8tpbG8k
But your basic options are multiplexing, using conjugated primaries (which you can often get cheaply for common housekeeping proteins), strip and reprobe, or cut and probe separately.
Alternatively, you can do a similar normalization to total protein levels, rather than to references. The protein levels can be measured with something like a reversible stain or stain-free technology. Such Total Protein Normalization (TNP) is increasingly being done as it becomes increasingly evident that housekeeping genes might not be as “constant” as we’ve thought.
Here’s a recent study that analyzed data on protein expression in different cell types, etc and found that our traditional go-tos weren’t very constant after all! Instead, they recommend some other ones
Lee, H. G., Jo, J., Hong, H. H., Kim, K. K., Park, J. K., Cho, S. J., & Park, C. (2016). State-of-the-art housekeeping proteins for quantitative western blotting: Revisiting the first draft of the human proteome. Proteomics, 16(13), 1863–1867. https://doi.org/10.1002/pmic.201500344
And here’s a review article that talks about variability in the ones we often use: Li, R., & Shen, Y. (2013). An old method facing a new challenge: re-visiting housekeeping proteins as internal reference control for neuroscience research. Life sciences, 92(13), 747–751. https://doi.org/10.1016/j.lfs.2013.02.014
When it comes to choosing housekeeping controls, you need to take other things into account. Even if the expression of a housekeeping gene/protein is expressed constantly in all the samples you’re trying to compare between, that’s not all that matters. Instead, you also need to make sure that it’s expressed at levels kinda sorta similar to the thing you are trying to measure. You need your housekeeping control to be in the linear range of detection where the amount of signal corresponds to the amount of control. If you are above saturation levels, you won’t be able to tell if there are differences and won’t be able to accurately compare. This can especially be an issue if your thing of interest is expressed at really low levels so you have to load a lot.
One more thing before you go, another term that you may want to know! Since I talked about “constitutive” I also wanted to mention a related term, “basal.” Basal refers to the sort of “baseline” or “normal” levels of something – a system’s “default.” It’s what you see under typical, unperturbed, conditions if you don’t do anything to activate or repress it.
We can use it to describe…
…how active enzymes or receptors are in the absence of activating stimuli (see the agonist etc. post for more on this: blog form: https://bit.ly/bindingpartners YouTube: https://youtu.be/sVcfAGH65y4)
Or, and particularly relevant for today’s post, how highly a gene is regularly expressed (relying only on “generic” transcription factors to get made)
I think that’s enough terms for today! Speaking of which, you might have noticed, I’m working on a sort of series of short posts/vids on tricky terminology. Especially terms that get used but not really explained. I’m working on putting together a page on my blog with them all.
And there’s already a glossary you can check out: https://thebumblingbiochemist.com/glossary/
Here’s more on qPCR: http://bit.ly/rtrtqpcrprimer
More on western blots: http://bit.ly/westernblotworkflow ; YouTube: https://youtu.be/Oun6_u6E090
A few helpful resources:
Novus, Loading Control Handbook: images.novusbio.comhttps://images.novusbio.com › …PDFLoading Control Handbook
Li-Cor, Normalizing to Correct for Technique Variability https://www.licor.com/bio/applications/quantitative-western-blots/normalization#why-normalize
BIO-RAD, Total Protein Normalizationhttps://www.bio-rad.com/en-us/applications-technologies/total-protein-normalization?ID=PODYJQRT8IG9
