Counting phosphorylations: one, two, many…

Jeremy Gunawardena’s lab just published a paper that should probably be required reading for anyone in the habit of attempting to measure the relative levels of phosphorylated proteins using Western blots (Prabakaran et al. 2011.  Comparative analysis of Erk phosphorylation suggests a mixed strategy for measuring phospho-form distributions.  Mol. Syst. Biol. 7 482).  If you are in that category, be warned: you will find this paper depressing.

What Prabakaran et al. wanted to do was to find a way of determining the pattern of phosphorylations on a protein.  They chose the simplest situation possible — Erk, a protein with just two phosphorylated sites — and set out to develop a reliable method for finding out how much of the protein was phosphorylated at only site 1, how much at only site 2, and how much on both sites.

Did you realize that with all our technology, we still can’t do this?  Many people don’t. Quantitative mass spectroscopy techniques have recently made it possible to get a number for how much of the protein is phosphorylated at site 1 or site 2, but that still doesn’t tell you the distribution of the phosphoforms.  Suppose you have a protein that looks like this:

XXXS1XXX[cleavage site]XXXS2XXX

where S1 and S2 are the sites of phosphorylation. The [cleavage site], obviously, is the point at which the enzyme you’re using to chop the protein into peptides to run it on the mass spec acts.  When you analyze your peptides, you will have no idea whether the XXX[P]S1XXX peptides you see come from a protein in which just S1 is phosphorylated, or a protein in which both S1 and S2 are phosphorylated.  So, if you see 50% [P]S1 and 50% [P]S2, you won’t know whether this reflects a situation in which both sites are phosphorylated independently (leading to a mixed population of proteins with only S1, only S2, and both sites phosphorylated) or a situation in which S2 is only phosphorylated after S1 (50% of the protein is phosphorylated on both sites, and 50% not at all).  This could easily be biologically important, don’t you think?

You can, of course, set up your digestions differently.  In the case of Erk, both of the phosphorylation sites are on a single peptide if you cut the protein with trypsin.  This doesn’t solve the problem, though, because now you can differentiate between a doubly-phosphorylated peptide and a singly-phosphorylated one, but you can’t tell the difference between a peptide phosphorylated only on S1 and one that is phosphorylated only on S2.  It’s all very annoying — and we’re only dealing here with 2 sites.  What about proteins like p53, which has 18 potential serine/threonine phosphorylation sites?

The most common way to try to get a handle on the distribution of phosphorylated species — the authors call them phosphoforms — has been to use Western blots, detecting the presence of each phosphoform with a different antibody.  Antibodies that are supposedly specific for the three different possible phosphoforms of Erk ([P]S1 only, [P]S2 only, and [P]S1[P]S2) are available, but it’s not clear how accurate these measurements really are.  Prabakaran et al. set out to compare the measurements from Westerns with three other methods: peptide mass spec (MS), protein MS (similar to peptide mass spec, but analyzing whole proteins instead of peptides), and NMR.

The good news first: the measurements from the three non-Western methods are pretty consistent with each other.  You can guess the bad news: the Westerns don’t agree with the other methods at all well.  For example, in one case where the antibodies claim that there is about twice as much of the unphosphorylated protein than the doubly-phosphorylated protein in a sample, the peptide-MS suggests that there is actually ~4x more of the doubly-phosphorylated protein than the unphosphorylated protein.  If we accept the MS/NMR numbers (which is a leap of faith, since there is no gold standard to which we can compare any of these measurements; but it’s three against one at this point), then the antibody measurements are not only wrong, but wildly wrong.  And the wrongness is of a peculiar sort: sometimes the Westerns agree with the MS and NMR methods that there is more A than B, but sometimes the two measurements flatly contradict each other.

Why is this?  The authors suggest two reasons.  The first comes from the mathematics of what we are assuming when we compare the intensities of the signals given by two antibodies and try to infer the relationship between the signals and the underlying reality.  The mapping between the strength of the signal an antibody gives you and the amount of its target in a sample is rarely exactly linear: instead, signal (y) is proportional to the amount of sample (x) raised to a power.  Let’s call it α; the better the antibody, the closer α is to 1.  So for antibody A, which tracks one of your phosphoforms, y is proportional to x(α).  For antibody B, which tracks another of your phosphoforms, the exponent will be different; let’s call it ß.  The ratio between α and ß will determine how accurately your antibodies track the relative amounts of phosphoforms in your sample.  If you’re lucky enough to have two antibodies with identical signal to sample profiles, so that α/ß=1, then you’ll get accurate relative quantitation (provided that the two phosphoforms don’t interfere with each other; see below).  If not, then the relative ratio of A’s target to B’s target may appear to flip in your sample, when in reality nothing much has changed. Prabakaran et al. measured the exponent for each of their antibodies, and found that the antibodies to single phosphoforms have pretty terrible exponents: 0.2 in one case, and 0.4 in the other.  So that’s one problem.

The other problem is that we assume that antibodies are magic, that they can detect a phosphorylation on a single site of a protein while completely ignoring everything else that’s going on around that site.  For example, we assume that the antibodies that bind to [P]S1 do so independent of whether S2 is phosphorylated or not.  This is probably unlikely, especially when the two sites are as close as they are in Erk; and indeed the authors see measurement differences between MS and Westerns that they can’t explain just by looking at differences in exponents.

So — having demolished the usefulness of Western blots, do Prabakaran et al. offer you anything in their place?  Yes and no.  Yes, it’s encouraging that peptide-MS, protein-MS and NMR all give results that are consistent with each other; and by combining peptide-MS with protein-MS you should be able to get most of the information you used to think you could get from Westerns.  Peptide-MS can measure the amount of each single-site phosphorylation, while protein-MS can tell you how much of the protein is singly or doubly phosphorylated.  What MS doesn’t automatically give you is the amount of protein in each singly-phosphorylated state: protein-MS can’t distinguish between [P]S1-protein and [P]S2-protein.  This problem becomes even more severe as the number of sites increases.  Perhaps this problem can eventually be solved by using enzymes that cleave in different places, producing peptides whose phosphoforms can be more easily distinguished by chromatography in peptide-MS.

Prabakaran et al. also point out an advantage of both protein-MS and NMR over Westerns: both methods allow you to see modifications you weren’t expecting.  Native human Erk protein has only two known phosphorylation sites, but it turns out that the tagged Xenopus protein used here has two additional serines that can be phosphorylated.  So instead of the “hydrogen atom” of multisite phosphorylation — just two sites — the authors found themselves studying something more akin to lithium.  Just another of biology’s amusing little traps for the unwary.

Prabakaran S, Everley RA, Landrieu I, Wieruszeski JM, Lippens G, Steen H, & Gunawardena J (2011). Comparative analysis of Erk phosphorylation suggests a mixed strategy for measuring phospho-form distributions. Molecular systems biology, 7 PMID: 21487401