Differential Phosphorylation in response to a wild-type and mutant protein kinase

Andrew aji23 at drexel.edu
Mon Jan 12 13:03:21 EST 2004

Thanks to both you and Emir for replying to my thread, I really do
appreciate the input I'm getting.

Emir - yes, it is common for a kinase to have multiple phosphorylation
sites on a single substrate protein; in addition, these sites can
vary, which makes it difficult to map specific sites.  This is
particularly the case when a protein has an exposed surface in which
the overall charge is important.  In such an instance, it isn't
important persay which of the S/T/Ys are Ped, just that the surface
charge is made more negative (our kinase appears to have just this
situation, by the way, and in addition to phosphorylating substrates,
it is itself a substrate of its kinase activity).  Whew.

I am still against the notion of using the phospho-specific antibodies
on the basis of what Tom said - that this will not yield a
differential result between wt and mutant kinase.  Kinases come in two
flavors - S/T and Y.  S/T kinases never phosphorylate Y and vice
versa.  In addition, an S/T can sometimes phosphorylate a S that was
mutated to a T, so again, using antibodies to address this question
has a very high probability of failing.

Toms idea of isoelectric focusing, a protocol I am not overly familiar
with, sounds promising, and I will investigate this further.  There
very well might be a differing pI.  A general detail, but one that may
represent a differential change!

As for the ES vs. MADLI-TOF, the reason MALDI is superior in this
specific sort of problem is that the upper limit of detected peptide
is supposedly infinite, wereas ES has a cutoff of roughly 30 kDa if I
recall correctly.  My peptides come in chunks greater than 30 KDa so
MALDI would be the way to go.  However, if there was someone around
this place who could run some ES samples for me I'd probably try it! 
(The other reason I'm not currently pursuing that is I'm using
radiolabeling, and I don't think people would appreciate a hot MS.)  -
oh, and it's a HPLC-MS, not MS-MS.  That would be a lot more money :)

My reactions are currently in cell culture.  In the realm of molecular
biology, in vivo generally means in cell culture and in vitro means in
the test tube.  I know MDs and other facets of biology use in vitro to
address things in cell culture and in vivo in the organism.  So, to
clarify, I am doing things in cell culture.  We are currently setting
up a cell-free system as well, but we are still working out the
details and it is very slow-going.

And no, I have no idea which residues are phosphorylated.

The Hunter protocol entails the following:  starting material =
radiolabeled protein.  Mine is immunoprecipitated out of labeled
cells.  The protein is run on a gel, transfered to n.c., exposed to
film and cut out.  the fragment is digested, lyophilized, and spotted
on a TLC plate.  THe first dimension is run horizontally and separates
the peptides by M/S (with highly + charged peptides running farther
than -).  The second dimension is a chromatography step, and is done
in a buffer that causes migration differences based on hydrophobicity,
with the more hydrophic peptides running higher.  So you are left with
a 2D surface in which the more phosphorylated something is, the lower
and to the left it runs.  My substrate maps are beautiful but
superimposible regardless of what I do to the damn things.  I know the
ratio of overexpressed proteins is not correctly and I am currently
trying to simply optimize the conditions.

> So, if i may ask, what makes you think the mutant kinase does have
> different behaviour?

We think the mutant kinase has different behavior based on the
phenotype it exhibits in cell culture; it is markedly different than
wildtype.  Yet the protein itself is still phosphorylated both in cell
free and in cells, indicating one of two things - it is still a
kinase, or it is strongly binding to a co-precipitating kinase.  So
far results are still inconclusive.

> One possibility might be that the mutant kinase is switching from one
> phosphorylation site to another that is in the same peptide, which i don't
> think you'd be able to detect by TLC (or single MS) - pSXXXS looks the
> same as SXXXpS. If that's the case, cutting the protein with a different
> protease might do the trick.

This is a huge caveat to the whole procedure and one I am painfully
aware of.  This could be the case now, with the data I have; I have to
further prove or disprove this.

> How about making constructs which express fragments (eg individual
> domains) of the substrate and doing the assay with each one? If the
> differentially phosphorylated residue is the only phosphorylated residue
> in its fragment, you should get a clear difference between the WT and
> mutant kinases (which you could detect by anti-pS/T WB, or radiophosphate
> labelling plus SDS-PAGE and autoradiography or scintillation counting).
> You could combine this with the point-mutation approach to suppress
> individual sites in each fragment; that would make the job quite a bit
> simpler.

This is an idea I was going to pursue if I clarified the actual
expression and phosphorylation problems I outlined above.  Excellent

Thanks for the input guys!


More information about the Methods mailing list